See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/267754594
Universal health coverage in emerging economies: Findings on health care
utilization by older adults in China, Ghana, India, Mexico, the Russian
Federation, and South Africa
Article  in  Global Health Action · November 2014
DOI: 10.3402/gha.v7.25314 · Source: PubMed
CITATIONS READS
26 119
12 authors, including:
Karl Peltzer Jennifer Anne Stewart Williams
Human Sciences Research Council (HSRC) Umeå University
684 PUBLICATIONS   8,839 CITATIONS    51 PUBLICATIONS   733 CITATIONS   
SEE PROFILE SEE PROFILE
Paul Kowal Joel Negin
Research Institute for Health Sciences The University of Sydney
236 PUBLICATIONS   2,677 CITATIONS    98 PUBLICATIONS   1,471 CITATIONS   
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Disability Grant for PLWHI in the Eastern Cape View project
SAGE Wellbeing in Older Persons Study (WOPS) HIV (SAGE-WOPS HIV) View project
All content following this page was uploaded by Jennifer Anne Stewart Williams on 10 November 2014.
The user has requested enhancement of the downloaded file.
Global Health Action 
ORIGINAL ARTICLE
Universal health coverage in emerging economies:
findings on health care utilization by older adults in
China, Ghana, India, Mexico, the Russian Federation,
and South Africa
Karl Peltzer1,2,3#*, Jennifer Stewart Williams4,5#, Paul Kowal5,6#,
Joel Negin7#, James Josh Snodgrass8, Alfred Yawson9, Nadia Minicuci9,10,
Liz Thiele11, Nancy Phaswana-Mafuya1,12, Richard Berko Biritwum9,
Nirmala Naidoo6,13 and Somnath Chatterji5,13 on behalf of the SAGE
Collaboration
1Human Sciences Research Council, Pretoria, South Africa; 2Department of Psychology, University of the
Free State, Bloemfontein, South Africa; 3ASEAN Institute for Health Development, Mahidol University,
Nakhon Pathom, Thailand; 4Department Public Health and Clinical Medicine Epidemiology, Global Health
Umeå University, Umeå, Sweden; 5Research Centre for Gender, Health & Ageing, University of
Newcastle, Newcastle, Australia; 6World Health Organization, SAGE, Genève, Switzerland; 7School of
Public Health, University of Sydney, Sydney, Australia; 8Department of Anthropology, University of
Oregon, Eugene, OR, USA; 9Department of Community Health, University of Ghana, Accra, Ghana;
10National Research Council, Institute of Neuroscience, Padova, Italy; 11Independent Consultant, Atlanta,
GA, USA; 12Office of the Deputy-Vice Chancellor, Nelson Mandela Metropolitan University, Port Elizabeth,
South Africa; 13World Health Organization, HIS/HSI, Genève, Switzerland
Background and objective: The achievement of universal health coverage (UHC) in emerging economies is a
high priority within the global community. This timely study uses standardized national population data
collected from adults aged 50 and older in China, Ghana, India, Mexico, the Russian Federation, and South
Africa. The objective is to describe health care utilization and measure association between inpatient and
outpatient service use and patient characteristics in these six low- and middle-income countries.
Design: Secondary analysis of data from the World Health Organization’s Study on global AGEing and adult
health Wave 1 was undertaken. Country samples are compared by socio-demographic characteristics, type of
health care, and reasons for use. Logistic regressions describe association between socio-demographic and
health factors and inpatient and outpatient service use.
Results: In the pooled multi-country sample of over 26,000 adults aged 50-plus, who reported getting health
care the last time it was needed, almost 80% of men and women received inpatient or outpatient care, or both.
Roughly 30% of men and women in the Russian Federation used inpatient services in the previous 3 years and
90% of men and women in India used outpatient services in the past year. In China, public hospitals were the
most frequently used service type for 52% of men and 51% of women. Multivariable regression showed
that, compared with men, women were less likely to use inpatient services and more likely to use outpatient
services. Respondents with two or more chronic conditions were almost three times as likely to use inpatient
services and twice as likely to use outpatient services compared with respondents with no reported chronic
conditions.
Conclusions: This study provides a basis for further investigation of country-specific responses to UHC.
Keywords: health care use utilization; low- and middle-income; universal coverage
Responsible Editor: Nawi Ng, Umeå University, Sweden.
*Correspondence to: Karl Peltzer, HIV/AIDS/STIs and TB Research Programme, Human Sciences Research
Council, Private Bag X41, Pretoria 0001, South Africa, Email: kpeltzer@hsrc.ac.za
Received: 27 June 2014; Revised: 6 September 2014; Accepted: 12 September 2014; Published: 31 October 2014
#These authors contributed equally to this work.
Global Health Action 2014. # 2014 Karl Peltzer et al. This is an Open Access article distributed under the terms of the Creative Commons CC-BY 4.0 License 1
(http://creativecommons.org/licenses/by/4.0/), allowing third parties to copy and redistribute the material in any medium or format and to remix, transform, and
build upon the material for any purpose, even commercially, provided the original work is properly cited and states its license.
Citation: Glob Health Action 2014, 7: 25314 - http://dx.doi.org/10.3402/gha.v7.25314
(page number not for citation purpose)
Karl Peltzer et al.
common and long-running concern across the the capacity of health care systems in LMICs to align
Aworld relates to the impact of increasing longevity with the UHC agenda.on health care utilization (1, 2). This is a problem, The World Health Organization’s Study on global
particularly for low- and middle-income countries (LMICs) AGEing and adult health (SAGE) Wave 1 provides im-
where 65% of the world’s population aged 60 and older portant standardized international data on health in older
currently resides, with this proportion expected to increase adults in six LMICs. The SAGE population represents
to 79% by 2050 (3). There are now strong ethical, political, 43% (700 million) of the global population aged 50
economic and health arguments being discussed and de- and older and 60% of those living in developing countries.
bated in global forums regarding universal health coverage The SAGE data allow comparative descriptions of na-
(UHC) (4). Complicating matters in LMICs is the fact that tional patterns of health care use in China, Ghana, India,
ill health is changing in character. Morbidity and mortal- Mexico, Russia, and South Africa. The objective of this
ity burdens from non-communicable diseases are exceed- study is to describe health care utilization and measure
ing burdens resulting from communicable diseases in an association between inpatient and outpatient service use
increasing number of developing countries (5). and patient characteristics in these six LMICs.
Although these demographic and epidemiological tran-
sitions may not have as large an impact on overall health Methods
care use and costs as once predicted, there is an expecta-
tion that policy-makers will ensure that health and social Study population
systems are adequately equipped to support aging popula- SAGE is a longitudinal study with nationally representa-
tions (6). The emergence of global interest in UHC may tive samples of adults in China (2008/10), Ghana (2008/09),
provide impetus to address these and other major policy India (2007/08), Mexico (2009/10), the Russian Federa-
challenges faced by governments in LMICs. For example, tion (2008/10), and South Africa (2007/08). All six SAGE
healthier aging cohorts and risk pooling under UHC may countries used stratified multistage cluster sampling stra-
provide a path for sustainable health service financing tegies with strata defined to ensure representation of a
and delivery (7). In the longer term, contemporary health range of living conditions and urban and rural localities in
care systems will need to provide at least a basic level of each country. Face-to-face structured interviews based on
care for the increasing numbers of older citizens, ensure standardized interviewer training procedures and survey
an equitable distribution of health care resources across instruments were used to collect household and individual
all age groups, undertake structural and organizational level data. One household questionnaire was completed
changes to manage chronic conditions, and improve the per household and all individuals aged 50-plus in selected
co-ordination of care within existing systems (1). households were invited to participate. Additional details
Health care utilization among older adults in both LMICs about SAGE are provided elsewhere (19).
and high-income countries is variously influenced by demo-
Variables
graphic, economic, and health status factors. Older women
The study population is described by: age groups (5059,
use primary care and community health services more than
6069, 7079, and 80-plus); sex (men and women); re-
older men (8, 9), and older men use inpatient services more
sidence (urban vs. rural), and ordered quintiles of rela-
often than older women (10). Studies in a number of LMICs
tive wealth. A random-effects probit model was used to
show higher inpatient utilization in urban compared with
estimate wealth levels based on asset ownership in each
rural areas (9, 11). In China, researchers identified rural
country (20). Estimates of asset-based wealth were applied
urban disparity in health service use with rural respondents to every household in each survey and used to establish
using physicians services less than urban respondents and quintiles of household wealth for each country. Quintile
hospitals more (12). Factors associated with health care one represents the lowest (poorest) fifth of respondents
utilization among older adults include older age (13), higher in terms of their asset-based wealth and the wealthiest fifth
education, living alone, worse self-perceived health (14), of respondents is represented in quintile five. A variable
chronic illness (15), and depressive symptoms (16), as well as indicating the number of possible chronic conditions was
the availability of accessible, affordable, quality health care included in the regressions. This variable was categorized
(17, 18). as - zero versus one versus two or more chronic conditions,
Evidence is necessary to inform the ongoing debate these being arthritis, depression, asthma, and diabetes.
about longevity and healthy aging in both developed and Self-assessed need for health care was ascertained by
developing countries. Although some data on health care the question: ‘The last time you needed health care,
utilization in older adults are available from LMICs, most did you get health care?’ Those who answered ‘yes’ were
of the research and policy attention has to date been in directed to further questions about their use of various
high-income countries. Sound evidence about health care types of inpatient and outpatient services, including long-
utilization by older adults in LMICs is necessary to gauge term care, home care, or the use of traditional healers.
2 Citation: Glob Health Action 2014, 7: 25314 - http://dx.doi.org/10.3402/gha.v7.25314
(page number not for citation purpose)
Universal health coverage in emerging economies
Different time-frames were used in questions on inpatient Multivariable logistic regressions separately describe
and outpatient care in order to capture sufficient incident statistical association between socio-demographic factors,
episodes and minimize recall bias (21). Inpatient use chronic conditions, country of residence, and inpatient
referred to a stay in a hospital and/or long-term care and outpatient service use (as dependent binary outcome
facility for at least one night in the previous 3 years. variables). The selection of covariates included in multi-
Outpatient use referred to ambulatory care received in the variable models was informed by the literature. P-values
previous 12 months excluding an overnight hospital stay. and variance inflation factors (VIF) are reported. All
analyses were carried out using STATA version 11.1
Data analysis (StataCorp, 2009).
Household-level and person-level analysis weights, based
on the selection probability at each stage of sampling Ethics
along with post stratification corrections, were applied SAGE has been approved by the World Health Organiza-
to produce nationally representative cohorts (19). Age and tion’s Ethical Review Committee. In addition, each partner
sex standardizations based on WHO’s World Standard organization implementing SAGE obtained ethical clear-
Population (22) were carried out to adjust for between- ance through their respective review bodies. Written in-
country age and sex differences. formed consent was obtained from all study participants.
This study presents weighted analyses of data self-
reported by adults aged 50 and older from China, Ghana, Results
India, Mexico, the Russian Federation, and South Africa. Table 1 gives summary statistics for the weighted study
Individual country and pooled multi-country populations samples: China 13,096, Ghana 4,300, India 6,545, Mexico
are described by age, sex, residence, and wealth quintiles 2,304, the Russian Federation 3,937, South Africa 3,817,
for China, Ghana, India, Mexico, the Russian Federation, and pooled 33,798. There were 133 records excluded from
and South Africa. Only records with complete data on the descriptive analysis because of missing data (61 from
these socio-demographic variables were included. Subse- China, 5 from Ghana, 38 from India, 4 from Mexico, 5
quent analyses were undertaken on sub-samples (indivi- from the Russian Federation, and 20 from South Africa).
dual country and pooled) in order to describe responses The percentage of men was low in the Russian Federation
to specific questions on health care use and need, types (39%) compared with China (50%), Ghana (53%), India
of services, and reasons for use. The denominators in- (51%), and Mexico (47%). The proportion of adults aged
clude only records with complete data for the questions of 80 and over was 10% in Ghana comparedwith 5% in China
interest. Imputation methods were not used in this study. and India, 9% in Mexico, 8% in the Russian Federation,
Table 1. Percent distribution of socio-demographic characteristics by countrya and pooledb, adults aged 50, SAGE Wave 1
China Ghana India Mexico Russian Fed South Africa Pooled
N 13,096 4,300 6,545 2,304 3,937 3,817 33,798
Age group % % % % % % %
5059 44.9 39.8 48.6 48.1 45.2 50.0 49.7
6069 31.8 27.5 30.9 25.6 24.6 30.6 28.6
7079 18.6 23.1 16.0 17.8 21.8 14.0 16.1
80 4.6 9.7 4.5 8.6 8.4 5.5 5.7
Sex
Men 49.8 52.5 51.0 46.8 38.9 44.0 48.8
Women 50.2 47.5 49.0 53.2 61.1 56.0 51.2
Residence
Urban 47.5 41.0 29.0 78.8 72.7 65.0 44.2
Rural 52.5 59.0 71.0 21.2 27.3 35.1 55.8
Wealth quintiles
Quintile 5 (highest wealth) 21.8 21.6 23.9 26.6 24.6 21.4 22.9
Quintile 4 23.4 20.7 19.6 16.6 20.5 19.8 21.9
Quintile 3 20.5 20.5 18.8 16.8 19.1 18.2 19.6
Quintile 2 18.1 19.1 19.5 24.7 19.6 19.9 18.8
Quintile 1 (lowest wealth) 16.3 18.2 18.2 15.3 16.2 20.7 16.9
aIndividual country survey weights applied. bPooled country survey weights applied. Russian FedRussian Federation. PooledPooled
country data.
Citation: Glob Health Action 2014, 7: 25314 - http://dx.doi.org/10.3402/gha.v7.25314 3
(page number not for citation purpose)
Karl Peltzer et al.
and 6% in South Africa. Most respondents lived in urban Utilization of health care: reason(s) for use
areas in Mexico (79%), the Russian Federation (73%), and Respondents who self-reported getting health care the last
South Africa (65%). Mexico had the highest proportion of time it was needed (the denominator) self-selected main
respondents in the wealthiest quintile (27%) and South reasons for their use of health care services from a list of
Africa had the highest proportion in the poorest quintile possible response options. Table 5 shows the top most
(21%). common reasons for respondents’ last inpatient overnight
In the pooled multi-country analysis, 5% of men and hospital stay in the previous 3 years, and last outpatient
4% of women reported that they did not get health care service in the previous year. Common reasons for inpatient
the last time it was needed (Table 2). Across the countries, service use included problems with the heart (20% in the
the proportions ranged from 1% in Mexico to 14% in
Russian Federation and 14% in China), generalized pain
Ghana.
in Ghana (17%), acute conditions (17% in India), and
Utilization of health care: type of facility communicable diseases in South Africa (13%). Common
Respondents who reported getting health care the last reasons given for outpatient service use were acute condi-
time it was needed selected the types of services that they tions (33% in China and 42% in India), generalized pain in
had used most frequently in the previous 3 years (Table 3). Ghana (23%), and high blood pressure in Mexico (17%),
Relative to the other countries, private clinics and health the Russian Federation (25%), and South Africa (39%).
care facilities were used frequently in India (by 38%
of men and 42% of women), as were private hospitals Factors associated with inpatient and outpatient
(by 20% of men and 16% of women). Public clinics and service use
health care facilities were used frequently in the Russian In the crude model (not shown) compared with China,
Federation (by 79% of men and also 79% of women) and respondents in the Russian Federation were 40% more
public hospitals were used frequently in China (by 52% likely to use inpatient services (OR1.4 95% CI: 1.1,1.7),
of men and 51% of women). The ‘other services’ category respondents in Ghana were 50% less likely to use inpa-
includes charity clinics and hospitals and traditional tient services (OR0.5 95% CI: 0.4,0.6), and respondents
healers, and these services were frequently used in Ghana in India (OR0.6 95% CI: 0.5,0.7) and Mexico (OR0.6
(by 27% of men and 25% of women). 95% CI: 0.4,0.8) were 40% less likely to use inpatient
services.
Utilization of health care: inpatient and outpatient
Table 6 shows the multivariable logistic regressions
services
for inpatient and outpatient use. Inpatient service use
Table 4 shows inpatient and outpatient service use by those
was 50% higher in the 7079 year age group (OR1.5
who reported getting health care the last time care was
95% CI: 1.3,1.7) compared with the 5059 year age group.
needed. In the pooled sample of over 26,000 older adults,
Women were significantly less likely to use inpatient
almost 80% received inpatient or outpatient care, or both.
Relative to the other countries, the use of inpatient services services than men (OR0.8 95% CI: 0.7,0.9). Respon-
in the previous 3 years was high in the Russian Federation dents in the lowest wealth quintile were 20% less likely
(28% for men and 30% for women). Past-year outpatient to use inpatient services compared with respondents in
service use was high in India with 87% of men and 90% the highest wealth quintile (OR0.8 95% CI: 0.7,1.0).
of women reporting use. In the multi-country sample, a Respondents with two or more chronic conditions were
similar proportion of men and women reported use of more than twice as likely to use inpatient services (OR2.8
inpatient services in the previous 3 years (21% vs. 19%) but 95% CI: 2.4,3.3) compared with respondents with no
a higher proportion of women reported use of outpatient recorded chronic conditions. The VIF (1.04) was within
services in the past year (72% vs. 68%). acceptable range for collinearity.
Table 2. Percent of men and women who self-reported not getting health care the last time it was needed, countrya and pooledb,
adults aged 50, SAGE Wave 1
China Ghana India Mexico Russian Fed South Africa Pooled
Total men N 4,951 2,001 2,886 854 1,072 1,272 12,881
Men (%) 5.9 14.2 2.1 1.1 8.9 1.5 4.9
Total women N 5,271 1,886 2,901 1,083 1,903 1,661 14,317
Women (%) 5.2 13.6 2.4 0.6 1.9 1.1 4.0
Denominators include only respondents who answered question on self-assessed health care. aIndividual country survey weights applied.
bPooled country survey weights applied. Russian FedRussian Federation. PooledPooled country data.
4 Citation: Glob Health Action 2014, 7: 25314 - http://dx.doi.org/10.3402/gha.v7.25314
(page number not for citation purpose)
Universal health coverage in emerging economies
Table 3. Types of health care services most frequently used in previous 3 years, by sex, countrya, and pooledb, adults aged 50,
SAGE Wave 1
China Ghana India Mexico Russian Fed South Africa Pooled
Men N 4,659 1,718 2,825 845 976 1,252 12,255
Women N 4,998 1,630 2,832 1,076 1,866 1,642 13,749
Private clinic, HC facility
Men% 19.8 10.1 37.6 32.5 2.1 32.0 25.2
Women% 20.2 7.6 41.7 24.0 2.5 28.0 26.0
Private hospital
Men% 1.0 7.0 19.7 3.8 0.1 4.1 7.0
Women% 1.4 6.3 16.2 8.6 0.2 1.6 5.8
Public clinic, HC facility
Men% 22.0 19.2 5.0 51.3 79.0 45.0 20.8
Women% 22.9 24.9 6.1 54.6 78.5 53.4 24.8
Public hospital
Men% 52.0 36.8 18.0 7.6 11.2 14.7 37.0
Women% 50.6 36.6 17.6 8.2 12.7 14.8 34.6
Other services not included
Men% 5.3 27.0 19.7 4.8 7.6 4.3 10.0
Women% 5.0 24.7 18.5 4.6 6.1 2.2 8.7
Private clinics include private health care facilities excluding hospitals. Public clinics include private health care facilities excluding
hospitals. Denominators include only respondents who self-reported getting HC the last time it was needed. aIndividual country survey
weights applied. bPooled country survey weights applied. HCHealth care. Russian FedRussian Federation. PooledPooled country
data.
In the crude model (not shown) with China as the 95% CI: 1.0,1.4) to use outpatient services than respon-
reference country, respondents in India were almost five dents in the 5059 year age group (Table 6). Respondents
times more likely to use outpatient services (OR4.8 95% in the lowest wealth quintile were 30% less likely to use
CI: 3.9,6.0), respondents in Ghana were 60% more likely to outpatient services compared with respondents in the
use outpatient services (OR1.6 95% CI: 1.4,2.0), and highest wealth quintile (OR0.7 95% CI: 0.5,0.8). Re-
respondents in Mexico were 60% less likely to use out- spondents with two or more chronic conditions were
patient services (OR0.4 95% CI: 0.3,0.6). almost twice as likely to use outpatient services (OR1.9
Women were 20% more likely to use outpatient services 95% CI: 1.6,2.3) compared with respondents with no
than men (OR1.2 95% CI: 1.1,1.3) and respondents in recorded chronic conditions. The VIF (1.11) was within
the 7079 year age group were 20% more likely (OR1.2 acceptable range for collinearity.
Table 4. Use of inpatient and outpatient services by sex, countrya and pooledb, adults aged 50, SAGE Wave 1
China Ghana India Mexico Russian Fed South Africa Pooled
Inpatient in the past 3 yearsc
Men N (%) 4,676 (24.1) 1,733 (14.1) 2,827 (15.6) 845 (17.6) 1,118 (27.6) 1,308 (12.1) 12,380 (20.7)
Women N (%) 5,015 (21.7) 1,637 (11.0) 2,833 (13.7) 1,076 (10.7) 2,047 (29.2) 1,683 (11.4) 13,904 (19.3)
Outpatient in the past yeard
Men N (%) 4,666 (60.5) 1,723 (70.1) 2,828 (87.3) 845 (36.3) 1,119 (60.8) 1,314 (55.9) 12,366 (68.1)
Women N (%) 5,010 (64.6) 1,645 (77.1) 2,834 (90.4) 1,076 (44.3) 2,047 (69.8) 1,684 (67.2) 13,894 (72.3)
Inpatient, outpatient, or bothe
Men N (%) 4,696 (71.8) 1,740 (75.1) 2,829 (90.4) 845 (47.9) 1,119 (71.0) 1,335 (60.5) 12,418 (76.5)
Women N (%) 5,030 (73.8) 1,646 (80.6) 2,835 (92.9) 1,076 (49.5) 2,047 (79.0) 1,722 (68.9) 13,937 (79.1)
Denominators include only respondents who self-reported getting HC the last time it was needed. aIndividual country survey weights
applied. bPooled country survey weights applied. Russian FedRussian Federation. PooledPooled country data. cRespondents who
used inpatient services in the past 3 years. dRespondents who used outpatient services in the past year. eRespondents who used either
inpatient services in the past 3 years, or outpatient services in past year, or both.
Citation: Glob Health Action 2014, 7: 25314 - http://dx.doi.org/10.3402/gha.v7.25314 5
(page number not for citation purpose)
Karl Peltzer et al.
Table 5. Top most common reasons for last inpatient overnight hospital stay in previous 3 years, last outpatient service in
previous year by countrya, adults aged 50, SAGE Wave 1
China Ghana India Mexico Russian Fed South Africa
Inpatient stays
Problems with Generalized pain Acute conditions Surgery (17.6%) Problems with Communicable
heart (14.0%) (16.9%) (17.2%) heart (20.0%) diseases (13.1%)
Surgery (10.1%) Surgery (13.9%) Surgery (13.2%) Problems with heart High blood Injury (13.1%)
(12.8%) pressure (17.4%)
High blood Communicable Problems with heart Cancer (11.1%) Surgery (14.0%) High blood pressure
pressure (6.4%) diseases (11.7%) (5.6%) (7.4%)
Injury (6.2%) Acute conditions Injury (6.6%) Diabetes (10.5%) Pain in joints (7.9%) Stroke (7.4%)
(11.0%)
Generalized pain High blood pressure Communicable Occupational Problems with Generalized pain
(5.2%) (6.5%) diseases (5.5%) (5.1%) breathing (5.0%) (7.0%)
Outpatient service
Acute conditions Generalized pain Acute conditions High blood pressure High blood High blood pressure
(33.1%) (23.0%) (41.5%) (17.2%) pressure (24.9%) (38.8%)
High blood Communicable Pain in joints (10.2%) Acute conditions Pain in joints Pain in joints (13.1%)
pressure diseases (13.6%) (14.0%) (12.7%)
(12.8%)
Pain in joints Pain in joints (12.6%) Generalized pain Diabetes (13.7%) Problems with Diabetes (10.3%)
(9.2%) (9.6%) heart (12.6%)
Generalized pain Acute conditions High blood pressure Nutritional (9.9%) Problems with Acute conditions
(5.3%) (12.5%) (6.7%) mouth (11.2%) (9.6%)
Problems with High blood pressure Problems with Pain in joints (5.7%) Acute conditions Problems with heart
heart (4.9%) (11.1%) breathing (3.4%) (7.6%) (3.5%)
Denominators include only respondents who self-reported getting health care the last time it was needed. Main reasons expressed as %
of all responses to these questions within each country. aIndividual country age survey weights applied. Russian FedRussian
Federation.
Discussion with a smaller proportion of out-of-pocket expenditure
Overall, the use of health care services in the 3 years prior (Mexico, South Africa, and the Russian Federation).
to SAGE interview varied by country. It is possible that Although less than 5% of older adults in the SAGE
differences in out-of-pocket payments influenced these countries self-reported not getting health care the last time
findings to some extent. When out-of pocket-payments it was needed, information on health care delivery systems,
are high, people often delay or defer accessing or using including out-of-pocket payments, doctor/patient ratios,
services even if they believe they need care (2325). Out- types and location of facilities, is needed to further in-
of-pocket payments account for a large share of total vestigate this finding. The increasing prevalence of chronic
health expenditure in LMICs (26). In 2010, out-of-pocket diseases among older adults in LMICs (2831) will impact
expenditure as a percentage of total health expenditure on access to affordable health care as countries move
was 35% in China, 28% in Ghana, 62% in India, 47% toward UHC.
in Mexico, 36% in the Russian Federation, and 7% in This study showed that the presence of one or more
South Africa (27). Between 2007 and 2012, price adjusted chronic health conditions was associated with higher out-
out-of-pocket expenditure increased by 54% in China, patient and inpatient service utilization. Studies in Mexico,
52% in the Russian Federation, 30% in Ghana, 24% in Ireland, Scotland, the United States, the Netherlands,
India, and 9% in South Africa. During the same period Switzerland, South Korea, and Germany reported similar
in Mexico, price adjusted out-of-pocket expenditure fell findings with multiple co-morbidity related to both
by 5% (27). The proportion of older adults in the SAGE higher utilization and costs for health care systems (11,
countries who reported not getting health care the last 13, 32).
time it was needed was higher in countries in which there The main reasons self-reported for the use of inpatient
was a higher proportion of out-of-pocket expenditure and outpatient services in the SAGE countries is con-
(Ghana and China) and relatively lower in countries sistent with evidence from the WHO (33). The African
6 Citation: Glob Health Action 2014, 7: 25314 - http://dx.doi.org/10.3402/gha.v7.25314
(page number not for citation purpose)
Universal health coverage in emerging economies
Table 6. Logistic regression of factors associated with inpatient and outpatient service use, adults aged 50, SAGE Wave 1
Inpatient use N26,177 Outpatient use N154
Odds ratio 95% CI Odds ratio 95% CI
Country (Reference: China)
Ghana 0.5 (0.4,0.6)*** 1.6 (1.4,2.0)***
India 0.5 (0.4,0.6)*** 4.4 (3.5,5.6)***
Mexico 0.5 (0.4,0.8)** 0.4 (0.3,0.6)***
Russian Federation 1.2 (0.4,0.8) 1.1 (0.8,1.5)
South Africa 0.4 (0.4,0.5)** 1.0 (0.8,1.5)
Sex (Reference: men) 1.0 1.0
Women 0.8 (0.7,0.9)*** 1.2 (1.1,1.3)***
Age group (Reference: 5059 years) 1.0 1.0
6069 1.1 (1.0,1,3) 1.0 (0.9,1.2)
7079 1.5 (1.3,1.7)*** 1.2 (1.0,1.4)*
80 1.2 (0.9,1.6) 1.1 (0.9,1.3)
Residence (Reference: urban) 1.0 1.0
Rural 1.1 (0.9,1.3) 1.4 (1.1,1.7)**
Wealth quintiles (Reference: highest wealth) 1.0 1.0
Quintile 4 1.0 (0.8,1.2) 1.0 (0.8,1.2)
Quintile 3 0.9 (0.8,1.1 0.9 (0.7,1.2)
Quintile 2 0.8 (0.7,1.0) 0.8 (0.6,1.0)*
Quintile 1 (lowest wealth) 0.8* 0.7 (0.5,0.8)**
Chronic conditions (Reference: none) 1.0 1.0
One chronic condition 1.8 (1.5,2.0)*** 1.6 (1.4,1.7)***
Two or more chronic conditions 2.8 (2.4,3.3)*** 1.9 (1.6,2.3)***
Denominator includes only respondents who received health care the last time it was needed. Inpatient use refers to the previous 3 years.
Outpatient use refers to the previous 12 months. Pooled country survey weights applied. ***pB0.01, **0.01BpB0.05, *0.05BpB0.1.
Variance Inflation Factor for inpatient model1.04. Variance inflation factor for outpatient model 1.11. CIconfidence interval.
countries continue to have a relatively high mortality Compared with older men, older women in the six
burden due to communicable diseases, while in other SAGE countries used inpatient services less often and
LMICs the burden is relatively high for heart problems, outpatient services more often. These results are generally
high blood pressure, surgery and generalized pain. consistent with evidence from studies conducted in high-
Research undertaken in high-income countries shows income countries, and also with results from a multi-
that health care utilization peaks at about 80 years of age country study which included nine LMICs (8, 9, 13). In
(34, 35). There was significantly higher inpatient and addition to gendered differences in service needs, social
outpatient utilization in the 7079 year age group in the relations and norms also influence access to health care
SAGE countries. (38). Other studies have found that men are more likely
Inpatient service use over the previous 3 years varied to use emergency services and be admitted to hospitals
from 11% (for men) in South Africa to 29% (for women) compared with women (8, 10). It is also possible that men
in the Russian Federation. Outpatient service use over have fewer preventive health visits and delay accessing
the previous year ranged from 36% (for men) in Mexico care when needed, which is consistent with men’s shorter
to 90% (for women) in India. A multi-country study of life expectancies compared with women’s. It has also been
people aged 65 and older showed that the proportion using suggested that, in terms of seeking health care, men and
health care services varied from 6 to 82% among sites in women react similarly to acute, life threatening, higher
Cuba, Mexico, Peru, Dominican Republic, Puerto Rico, severity illnesses, but differently to chronic, less severe
Venezuela, China, India, and Nigeria (9). Two Brazilian conditions (13).
studies, which investigated the use of primary care services Socioeconomic inequalities in health care use exist in
by adults aged 60 and older in the previous 6 months, most countries and in general, wealthier people are more
found that health care utilization ranged from 45 to 50% likely to use health care services. The findings from this
(15, 36). In a large nationally representative study con- study support other evidence of positive association
ducted in China, 30% of adults aged 60 and older reported between wealth and use of health care in rich and poor
no physician visits in the past year (37). countries (9, 38). Generally, economic factors, such as
Citation: Glob Health Action 2014, 7: 25314 - http://dx.doi.org/10.3402/gha.v7.25314 7
(page number not for citation purpose)
Karl Peltzer et al.
higher economic or wealth status, financial support, and care access and need, both met and unmet, taking into
health insurance, are associated with health care utiliza- account factors such as facilities, structures, systems of
tion among older adults (10, 13, 39). Older adults in lower operation, and the mix of public/private investment.
income countries may not have access to financial support
or pensions, making them more economically vulnerable Authors’ contributions
(9, 40). PK, SC, and NN were the originators of this paper, as
part of a set of initial publications for SAGE Wave 1. KP,
Strengths and limitations PK, and JSW drafted the initial paper. JSW and KP
The results should be interpreted with caution. Self-reported undertook the statistical analyses. JSW managed the
‘need’ may not represent true clinical need, many illnesses editing and finalization of the manuscript with input
and chronic conditions are undiagnosed, and perceptions of from JN, PK, and KP. All authors contributed to the
health care need may be different in the SAGE countries. final manuscript.
Health care utilization depends on a range of factors
that relate both to health care financing and deliv- Acknowledgements
ery systems and to individuals themselves. We did not
attempt to make adjustments for differences in health care We thank the respondents in each country for their continued
delivery and financing in each of the SAGE countries. contributions. We are grateful for the contributions of the country
primary investigators and their respective survey teams. The study
We acknowledge, however, that on the supply side, the on global AGEing and adult health is supported by the US National
structure and organization of health coverage and the Institute on Aging’s Division of Behavioral and Social Science
location, mix, and types of public and private providers Research (BSR) through Interagency Agreements (OGHA 040347
are all important factors, while on the receiving side, 85; YA1323-08-CN-0020; Y1-AG-1005-01) with WHO. Financial
patients’ age, mental health, functioning, socioeconomic and/or in-kind support has come from the governments of China and
South Africa (National Department of Health) to their respective
status, insurance status, proximity to care, perceptions of
national studies. The Shanghai CDC contributed financial and in-
the quality and benefits of care, and levels of health kind support for SAGE China. JSW was supported by the University
literacy can influence decisions to seek care (41). of Newcastle Australia Research Centre for Gender, Health and
We were unable to report the extent to which cultural, Ageing (early development) and subsequently by Umeå Centre for
contextual, economic, and structural factors, including Global Health Research, with support from FAS, the Swedish
health insurance, may have differently impacted the re- Council for Working Life and Social Research (Grant No. 2006-
1512).
porting of unmet need in this dataset. Most of the
research in this area has been undertaken in high-income
countries. Unmet need is an important indicator of health Conflict of interest and funding
care access. These findings suggest a need for further
The authors have not received any funding or benefits from
research on unmet need in LMICs, particularly in light of
industry or elsewhere to conduct this study.
the UHC agenda.
A major strength of this study is the use of standardized
questionnaires individually administered by trained inter- Financial disclosure
viewers to large representative samples of adult popula- None of the donor agencies had a role in data collection and
tions in LMICs from different geographic regions of the analysis, decision to publish, or preparation of the manuscript.
world. Questionnaires were translated into local languages
and all interviews were administered one-on-one in cul- References
turally appropriate settings. However, the cross-sectional
nature of the Wave 1 data presents limitations with regard 1. Rechel B, Doyle Y, Grundy E, McKee M. How can health systems
respond to population ageing. Copenhagen: WHO Regional Office
to inferring causality. When SAGE panel data become
for Europe; 2009.
available for Wave 2 and beyond, it will be possible to 2. Bloom DE, Mahal A, Rosenberg L. Design and operation of
investigate changes in health and health care utilization health systems in developing countries. Chapter 13. Global
over time and also compare trends in use in different population ageing: Peril or promise? Geneva: World Economic
regions. Forum; 2012, p. 148.
3. United Nations. World population ageing 2013; New York:
United Nations; 2013.
Conclusions 4. The struggle for universal health coverage [editorial]. Lancet
Both rich and poor countries require practical guidance on 2012; 380: 859.
ways to finance health systems and move toward UHC. 5. Salomon JA, Wang H, Freeman MK, Vos T, Flaxman AD,
No single policy mix will suit all countries. The findings Lopez AD, et al. Healthy life expectancy for 187 countries,
19902010: a systematic analysis for the Global Burden Disease
add to current understanding of patterns of service
Study 2010. Lancet 2012; 380: 214462.
utilization among adults aged 50 and older in six LMICs. 6. Suhrcke M, Fumagalli E, Hancock R. Is there a wealth dividend
This paper provides a basis for further research into health of aging societies? Public Health Rev 2010; 32: 377400.
8 Citation: Glob Health Action 2014, 7: 25314 - http://dx.doi.org/10.3402/gha.v7.25314
(page number not for citation purpose)
Universal health coverage in emerging economies
7. Bristol N. Global health action toward universal health cover- effects on utilization and out-of-pocket expenditure: evidence
age. Washington, DC: Center for Strategic and International from Vietnam. Int J Equity Health 2009; 8: 17.
Studies; 2014. 25. Galárraga O, Sosa-Rubı́ SG, Salinas-Rodriguez A, Sesma-
8. Redondo-Sendino A, Guallar-Castillón P, Banegas JR, Vázquez S. Health insurance for the poor: impact on cata-
Rodrı́guez-Artalejo F. Gender differences in the utilization of strophic and out-of-pocket health expenditures in Mexico. Eur J
health-care services among the older adult population of Spain. Health Econ 2010; 11: 43747.
BMC Public Health 2006; 6: 9. 26. World Health Organization (2010). The World Health Report 
9. Albanese E, Liu Z, Acosta D, Guerra M, Huang Y, Jacob KS, health systems financing: the path to universal coverage.
et al. Equity in the delivery of community healthcare to older Geneva: WHO.
people: findings from 10/66 Dementia Research Group cross- 27. World Health Organization National Accounts Indicators
sectional surveys in Latin America, China, India and Nigeria. [database on the Internet]. WHO; 2014. Available from: http://
BMC Health Serv Res 2011; 11: 11. apps.who.int/nha/database/Home/Index/en [cited 3 December
10. Roy K, Chaudhuri A. Influence of socioeconomic status, wealth 2013].
and financial empowerment on gender differences in health and 28. Yang W, Lu J, Weng J, Jia W, Ji L, Xiao J, et al. Prevalence of
healthcare utilization in later life: evidence from India. Soc Sci diabetes among men and women in China. N Engl J Med 2010;
Med 2008; 66: 195163. 362: 1090101.
11. Salinas J, Al Snih S, Markides K, Ray LA, Angel RJ. The 29. Roberts B, Stickley A, Balabanova D, Haerpfer C, McKee M.
rural  urban divide: health services utilization among older The persistence of irregular treatment of hypertension in the
Mexicans in Mexico. J Rural Health 2010; 26: 33341. former Soviet Union. J Epidemiol Community Health 2012; 66:
12. Liu M, Zhang Q, Lu M, Kwon C, Quan H. Rural and urban 107982.
disparity in health service utilisation in China. Med Care 2007; 30. Gao Y, Chen G, Tian H, Lin L, Lu L, Weng J, et al. Prevalence
45: 6006. of hypertension in China: a cross-sectional study. PLoS One
13. Wong R, Dı́az JJ. Health care utilization among older Mex- 2013; 8: e65938. doi: 10.1371/journal.pone.0065938.
icans: health and socioeconomic inequalities. Salud Publica 31. Addo J, Agyemang C, Smeeth L, de-Graft Aikins A, Edusei
Mex 2007; 49(Suppl 4): S50514. A K, Ogededegbe O. A review of population-based studies of
14. Guerra HL, Firmo JO, Uchoa E, Lima-Costa MF. The Bambuı́ hypertension in Ghana. Ghana Med J 2012; 46(Suppl 2): 411.
Health and Aging Study (BHAS): factors associated with 32. Al Snih S, Markides KS, Ray LA, Freeman JL, Ostir GV,
hospitalization of the elderly. Cad Saude Publica 2001; 17: Goodwin JS. Predictors of healthcare utilization among older
134556. Mexican Americans. Ethn Dis 2006; 16: 6406.
15. Paskulin LM, Valer DB, Vianna LA. Use and access of the 33. World Health Organization (2009). Global health risks. Mor-
elderly to primary health care services in Porto Alegre (RS, tality and burden of disease attributable to selected major risks.
Brasil). Cien Saude Colet 2011; 16: 293544. Geneva: WHO.
16. Peytremann-Bridevaux I, Voellinger R, Santos-Eggimann B. 34. Mcgrail K, Green B, Barer ML, Evans RG, Hertzman C. Age,
Healthcare and preventive services utilization of elderly Europeans costs of acute and long-term care and proximity to death:
with depressive symptoms. J Affect Disord 2008; 105: 24752. evidence for 19871988 and 19941995 in British Columbia.
17. Lloyd-Sherlock P. Population ageing in developed and develop- Age Ageing 2000; 29: 24953.
ing regions: implications for health policy. Soc Sci Med 2000; 35. Hessler RM, Eriksson BG, Dey D, Steen G, Sundh V, Steen B.
51: 88795. The compression of morbidity debate in aging: an empirical
18. Fenton JJ, Jerant AF, Bertaki KD, Franks P. The cost of test using the gerontological and geriatric population studies
satisfaction. A national study of patient satisfaction, health care in Göteborg, Sweden (H70). Arch Gerontol Geriatr 2003; 37:
utilization, expenditures, and mortality. Arch Intern Med 2012; 21322.
172: 40511. 36. Rodrigues MAP, Facchini LA, Piccini RX, Tomasi E, ThuméI
19. Kowal P, Chatterji S, Naidoo N, Biritwum R, Wu F. Data E, Silveira DS, et al. Use of primary care services by elderly
resource profile: the World Health Organization Study on global people with chronic conditions, Brazil. Rev Saúde Pública 2009;
AGEing and adult health (SAGE). Int J Epidemiol 2012; 41: 43: 9.
163949. 37. Li Y, Chi I. Correlates of physician visits among older adults in
20. Ferguson B, Murray CL, Tandon A, Gakidou E. Estimating China: the effects of family support. J Ageing Health 2011; 23:
permanent income using asset and indicator variables. In: 93353.
Murray CL, Evans DB, eds. Health systems performance 38. Rodin D, Stirbu I, Ekholm O, Dzurova D, Costa G,
assessment debates, methods and empiricism. Geneva: World Mackenbach JP, et al. Educational inequalities in blood
Health Organization; 2003, pp. 74761. pressure and cholesterol screening in nine European countries.
21. Clarke P, Fiebig DG, Gerdtham U-G. Optimal recall length in J Epidemiol Community Health 2012; 66: 10505.
survey design. J Health Econ 2008; 27: 12758. 39. Li Y, Chi I, Zhang K, Guo P. Comparison of health services use
22. Ahmad OB, Boschi-Pinto C, Lopez AD, Murray CJL, Lozano by Chinese urban and rural older adults in Yunnan province.
R, Inoue M. Age standardization of rates: a new WHO Geriatr Gerontol Int 2006; 6: 2609.
standard. Geneva: World Health Organization; 2001. 40. Kuo RN, Lai M-S. The influence of socio-economic status and
23. Wagstaff A, Lindelow M. Can insurance increase financial risk? multimorbidity patterns on healthcare costs: a 6-year follow-up
The curious case of health insurance in China. J Health Econ under a universal healthcare system. Int J Equity in Health
2008; 27: 9901005. 2013; 12: 11.
24. Axelson H, Bales S, Minh PD, Ekman B, Gerdtham UG. 41. Allin S, Masseria C. Unmet need as an indicator of health care
Health financing for the poor produces promising short-term access. Eurohealth 2009; 15: 79.
Citation: Glob Health Action 2014, 7: 25314 - http://dx.doi.org/10.3402/gha.v7.25314 9
(page number not for citation purpose)
View publication stats