Cross-country Evaluation of Health Care Systems Performance on Mortality Indices
Explaining the impact of health care systems in determining variations in health status across countries has a significant role in health policy. This study explored the effects of variations in the volume of health care and in certain characteristics of health systems on mortality indices across 20 MENA (Middle East and North Africa) countries. Latent class regression model was used in a panel structure study for classification of 20 Middle East and North Africa countries, over the period 1985-2005, based on health systems performance on mortality indices and explore the most important determinants of between country variations in health care systems efficiency. The results suggest that among a wide range of health system inputs, physicians number, education, level of immunization coverage and political stability are the most important factors that significantly affect between-country variations in terms of the effectiveness of health systems in improving mortality indices. One of the most important challenges in public health systems in MENA region is distinction between services and function that mainly steams from weak performance by governments. Trying to optimizing in management and monitoring of services, directing health expenditures toward cost-effective programs and reducing the gaps in health coverage can effectively improve health systems functioning in this region.
Received: October 31, 2011;
Accepted: December 29, 2011;
Published: January 18, 2012
For much of the last three decades, many countries, particularly developing countries, have introduced reforms aimed at improving the performance of health system. Among them are Middle East and North African (MENA) countries.
The MENA region is diverse, covering over 20 countries from the Western Sahara
to Iran. MENA has achieved significant improvements in health status in which
over the past three decades, infant and child mortality rates have fallen, health
service delivery has improved and overall life expectancy has risen. However,
given these remarkable progress, MENA countries still face a number of obstacles
in achieving efficient, equitable and sustainable health services for their
populations. Health disparities are also a challenge both across the region
and within each country (Pierre-Louis et al., 2004).
Therefore, the study of factors behind the cross-country differences in health
care systems efficiency would be necessary.
Evaluation of health care systems performance can be done by measuring the achievements of health systems across a range of objectives. The results of such evaluation correspond to the fundamental needs in health policy for policy makers. In comparative studies related to health systems performance between countries or organizations of particular interest is whether and how efficiency of systems vary and how they use different effective resources to improve health outcomes.
Hitiris and Posnett (1992), Grubaugh
and Santerre (1994), Or (2001) and Or
et al. (2005) showed that exists a positive relationship between
population health status and the input measures such as the number of physicians,
access to health facilities (e.g., the number of physician, medical centers,
hospital beds) total payments in health systems and the level of real resources
used by health system (Hitiris and Posnett, 1992; Grubaugh
and Santerre, 1994; Or, 2001, Or
et al., 2005).
The most comprehensive study about the influence of socio-economic factors
on infant and child mortality has been done through framework (Mosely
and Chen, 1984). These factors operate at community, household or individual
levels through five proximate determinants (Behm, 1991;
Sastry, 1997; Agha, 2000). For
example outside of medical factors, education appears to be another important
determinant of health status. It seems to determine many of decisions, which
affect the quality of life, ability to select a healthy diet, avoiding unhealthy
habits, efficient use of medical care and so on (Or et
al., 2005; Valkonen, 1989; Madise
et al., 1999; Macinko et al., 2006).
The framework developed by WHO for measuring and assessing health systems performance
specifies three intrinsic goals of the health system: Improving the health status
of population, equity in access to health care facilities, quality of health
facilities provision and financial protection (WHO, 2000).
Investigation about the impact of these factors on health status may help to
provide guidelines for health policy makers.
According to directed goals by WHO for health systems, it can be seen that the scope of health is enormous implicitly suggesting that collecting and compiling of related data and information would not be an easy task to do. Therefore, study and estimation of health systems performance and comparison of their effects on health status among the countries will not be straightforward. This study requires a multidisciplinary approach to distinguish relative impact of these factors.
In one hand, considering the relevant health outcomes of population can be
considered as the first step to assess and compare the countries performance
in health sector. In other hand, wide range and complexity of coherence scope
of health and health systems has led to development of composite measures of
health system performance. The rationale for developing a composite measure
is that no single metric can capture the concept of system performance. Instead,
each of the component dimensions of performance is measured and then various
components are combined into a single measure of whole system performance (Smith,
2002). Accordingly, this study using a suitable modeling approach i.e.,
"latent class regression" tries to provide a comprehensive indication of health
status through providing a composite indicator (as dependent variable) based
on three mortality indices i.e., life expectancy at birth (LEB),
Infant Mortality Rate (IMR) and under five Mortality Rate
(U5MR) in order to investigate the impacts of a wide range of medical
and non-medical effective factors on health status over 1985-2005 for Middle
East and North African countries. We concentrating on one aspect of health outcomes
i.e., mortality indices". These indices among other indicators related
to health status, especially at the national and international level provide
reliable information to describe the population health status. In addition,
they are readily available over a considerable time span.
Although, many studies on evaluating the national and cross-national level
health systems performance have been done. However, such researches in the developing
countries in this field are rare (Fuse and Crenshaw, 2006;
McGuire, 2006). On the other hand, Actual knowledge
on the determinants of health outcomes of populations, at country level and
in the scope of comparative researches related to health systems efficiency,
suffers from the partiality of the models and frameworks used in the empirical
Present study differs from much of the previous studies in some important aspects. To our knowledge, this is the first time that such a study, particularly, by applying of this modeling framework, as an alternative statistical modeling approach for health systems evaluation, in MENA region is done. In addition, the present study using a large sample (through panel data) and providing some methodological and conceptual controls through appropriate statistical model and introducing some important predictors, according to conditions of study countries, has tried to provide some alternative evidences of a statistically significant of determinants of health systems performance.
MATERIALS AND METHODS
Statistical method: Each measure of health status at national level
over a time span gives a different picture of health status that has been improved
over the time within the countries and may result in somewhat different ranking
of the health across the countries at any point in time. Any modeling approach
to explain or analyze of the performance of health systems will need to consider
these differences across the various measures of health outcomes both in trend
over time and in variation across the countries. Latent Class (LC) modeling
is one of the flexible approaches for analyzing the data with such a structure
(Deb and Trivedi, 2002; Jimenez-Martin
et al., 2002).
It is well known that LC models can be used to combine the information contained
in multiple observed outcome variables. In LC analysis, a discrete unobserved
variable with K categories is assumed to exist that affects the, usually multiple,
responses of subjects. Each subject is assumed to belong to one of these K latent
classes or mixture components.
In the present study, latent class regression is applied to modeling the cross-country heterogeneity in a flexible way that assumes no distribution for the unobserved country effects. Furthermore, it allows the distinction between latent classes of countries, which can vary as to the expected level of health status as well as to the responses to the covariates considered. Therefore, it is possible to assess to which extent the effects of the health system inputs (explanatory variables) considered in this study vary across different groups of countries (i.e., class specific covariates coefficients).
Variables and data
Outcome variables: For having a comprehensive perspective and may relieve some particular problems for a society, it may be useful to examine a range of measures rather than just one aspect of mortality in order to cover a wide range of population. Accordingly, in the present research, the outcomes of the health systems are restricted to three mortality indices including: LEB, IMR and U5MR within the years 1985-2005. They are commonly used for the comparison of health care systems, monitoring and designing of population health programs.
These variables play the role of the observed dependent variables in latent class regression modeling. Table 1 present the patterns of change in the considered mortality indices over the study period for MENA countries.
Predictor variables: An important problem in cross-national comparative studies on health systems performance, especially in developing countries, is availability and quality of data that affects the precise and robustness of statistical analysis. There are many potential determinants (medical and non-medical factors) of mortality which could be incorporated. An attempt was made in our study to fitting the most reliable predictor variables. The choice of predictor variables was done, among the most relevant available variables, based on economic theory, empirical evidence from earlier works in this area and considering the MENAs specific conditions. The predictors for entering them in the model are as follows:
||Number of physicians per 100,000 population, total health
expenditures (as a percentage of GDP) converted to international prices
for comparative purposes, immunization coverage for six target disease for
WHOs EPI program (i.e., measles, diphtheria, pertussis, tetanus, tuberculosis
and polio) as medical factors
||As non-medical factors, GDP per capita, GINI coefficient and rate of access
to safe water and sanitations were introduced in fitted model. In order
to determine the impact of socio-economic differences across countries as
an important component of the environmental effective factors on the health
outcomes, in the present study, GDP per capita, Gini coefficient and education
that play key role in these differences were used. political-structural
mechanisms and their interaction with national policies and institutions
can play effective role on health outcomes. In the present study, due to
the geopolitical conditions of Middle East and North African countries,
structural and political stability was introduced as a predictor.
Descriptive statistics including means and standard deviations for dependent
and independent variables are presented in Table 2
Data sources: Due to lack of a comprehensive data source, specially, for the study countries in the present research, the data was gathered from several sources including: United Nations Statistics Division, WHOs mortality database, United Nations site for the MDG Indicators, UNICEF Statistics Division, the WORLD Bank Group (HNP Stats), Joint Monitoring Program for Water supply and Sanitation (WHO-UNICEF), Statistical, Economics and Social Research and Training Center for Islamic Countries (SESRETCIC), Country Analytical Website (CAW) countries' specific website and year books.
Descriptive results: Table 1 displays the trend of mortality indices in MENA during the last 25 years, also Table 1 shows health status of MENA countries comparing to the developed, developing, least developed countries and world.
Table 2 shows the means and standard deviations of dependent
and independent variables during the 1985-2005 for MENA countries.
|| Comparison of MENA's mortality indicators with other regions
|*years **Per 1000 live births
||Mean and standard deviation of the dependent and independent
variables over the study period
|*%population-source: Joint monitoring program for water supply
and sanitation ,WHO-UNICEF, www.wssinfo.org,
**Gross domestic product based on purchasing-power-parity (PPP) per capita
||The class specific mean values of mortality indices and the
size of classes
One notable point in these results is the high standard deviations of the variables;
it can be attributed to the wide discrepancies of countries in terms of studied
Analytical results: One of the most relevant applications of latent class analysis is clustering or constructing typologies based on the observed variables.In the present research, models with 1 to 5 classes were fitted and the 3-class model turned out to performs best in terms of the BIC criterion.
The results of the fitted latent class regression model to the data at hand for classification of MENA countries and the effects of covariates are presented in Table 3 and 4. In addition, Table 5 shows countries included in each constructed class.
The estimated size of each latent class in the 3-class model are 28.6, 47.6
and 23.8. The mean values of mortality indicators in each class are displayed
in Table 3. Each class is named and its context (i.e., low,
middle and high) can be interpreted based on the mean values of mortality indices.
For example, class 1 contains the countries with low life expectancy and high
infant and under five mortality rates. Accordingly, this class has the worst
conditions among the three constructed classes in terms of studied mortality
indices (Table 3, 5).
In latent class regression model, the class membership probabilities for subjects or individuals (here, countries) are modeled as a function of covariates. As a result the class membership can be modeled as a Multinomial Logit. The presence of the covariates improves the prediction of latent classes as well as provides class specific covariates coefficients.
In the fitted model, similar as multinomial logistic regression, the Class 1 was considered as a reference category. Therefore, the covariates coefficients can be interpreted as "log odds ratios" or Odds Ratio (OR) belonging to Classes 2 and 3 rather than Class 1 in terms of one unit change in the related covariate adjusted for the values of other covariates.
The investigation of covariates coefficients reveals that the physician numbers, education, immunization coverage and political stability had the highest log odds of being a country in Classes 2 and 3 rather than Class 1. For example, the effect of one unit increase in physician numbers in terms of log odds ratio for belonging a country in Classes 2 and 3 rather than Class 1 is 1.061 and 0.989, respectively.
Our findings indicated that education is the second important effective factor on health status in MENA. As Table 4 shows the effect of education on being a country in Classes of 2 and 3 rather than Class 1 is similar.
The effect of other covariates can be similarly interpreted.
We included a control term for years of observation in fitted regression model.
This accomplishes two important objectives: (1) It controls for occurred improvements
in health outcomes that not captured by changes in other entered covariates
and (2) it prevents against a artificial association between trending variables.
|| The estimated class specific covariates (health systems inputs)
|| Classification of countries based on posterior probability
Our investigation based on fitting different models showed that the effect
of time (t = 1, 2, 3, 4, 5) was linear and class dependent. As Table
5 shows, the effect of time in Class 2 compared with Class 3 was more statistically
significant. This result indicates that the Class 2 has experienced higher growth
rate over 1985-2005 in terms of improvement in mortality indicators compared
with countries included in Class 3. The justification is that during the study
period, countries starting with high life expectancy and low IMR and U5MR levels
(third class countries) tended to improve less (maintained and optimized their
status) in terms of mentioned indicators than countries starting with low life
expectancy and high IMR and U5MR. Accordingly, the growth rate for countries
belong to Class 3 comparing to the countries being in Class 2, was slower over
this period. This can also be considered as a meaningful implication on the
priority of the effects of most of considered covariates (Health inputs) on
health status in Class 2 compared with Class 3 within the years 1985-2005.
Classification of countries: In latent class regression, the posterior probability of a subjects membership in each class can be computed and each subject (here, countries) is traditionally assigned to a group with the largest probability. Table 5 presents the results of assignment of each country to its most relevant class.
This research, as the first attempt on evaluating of the country-level determinants of population mortality indices over a time span in MENA countries, concluded that the cross-country variations in health status are substantially attributable to medical care facilities (i.e., physicians number, immunization coverage) education, income inequality, political conditions and partially to GDP and health expenditures.
In the present research, the most important effective factor is number of physicians.
In many studies, physicians role on cross-country variation in evaluation
of health systems, particularly in developed countries and somewhat in developing
countries has been investigated and its importance has been approved (Robst
and Graham, 1997; Robst, 2001; Nixon
and Ulmann, 2006; Papageorgiou et al., 2004).
The second effective predictor is level of education Literacy often has positive
effects on individuals viewpoint and quality of life (through choice of
job, ability to select a healthy diet, avoiding unhealthy habit, effective use
of medical care and
) therefore, it can play an effective role on health
status. Furthermore, parents' level of education has important implication for
child survival. In particular, mothers education level my increase their
productivity in child care (preventive and curative health care) (O'Hara,
1980; Behrman and Wolfe, 1984; Kassouf
and Senauer, 1996; Zakir and Wunnava, 1999; Currie
and Moretti, 2003).
Regarding to the effect of immunization coverage we can see the coefficient
value for this covariate in second class is relatively larger than third class,
indicating that vaccination coverage could have a larger impact on infant and
particularly under five mortality in countries being in the second class. This
results, might be attributed to higher and stabilized coverage level for countries
included in third class compared with the countries being in second class; consequently
marginal health benefits from expanding of vaccination coverage are likely to
be more significant in later countries. On the other hand, the implementation
of immunization programs in countries included in first class, as reference
class, was mainly affected by instability in their structural and political
Income as a proxy for standard living leads to improving in public health infrastructures
(safe water and sanitation, better nutrition ,better housing and ability to
paying for health care) particularly through increasing the capability of governments
and other policy makers to supply of the health care (Cutler
et al., 2006). The obtained results by some researchers such as Casterline
et al. (1989) on weak effect of income on infant and chilled mortality
can be considered as implications of less importance of this factor in affecting
the health status particularly in poor countries (Heerink,
1994; Casterline et al., 1989). Other researchers
such as Cutler et al. (2006), Preston
(1980), Caldwell (1986), Fogel
(1994) and Palloni and Hill (1997) obtained similar
results on weak effect of GDP on mortality (particularly on LEB). Most of these
researchers believe that the operation of income on mortality indicators is
through affecting the other effective parts of health systems such as development
of public health infrastructure and improvement of nutrition.
Although, health expenditures are major inputs of health systems however, the
effect this factor similar as GDP was weak on mortality indices especially in
Class 2. The previous studies in this subject area have failed to identify a
strong and consistent relationship between health expenditures and health outcomes
(after controlling for other covariates). For example, Burnside
and Dollar (1998) and Hill and Pebley (1989) found
no significant relationship between health expenditures and change in infant
mortality rate in low-income countries. Similarly, a number of other studies
found that the effect of health expenditure on health status (infant and child
mortality) is either small or statistically insignificant (Kim
and Moody, 1992; Musgrove, 1996; Filmer
and Pritchett, 1997, 1999). However, Gupta
et al. (1999) using data of 50 developing and transition countries
showed that health expenditures reduce child mortality (Gupta
et al., 1999). They also provided further evidences from 70 countries
on the importance of public spending on health status in low-income countries
than higher-income countries (Gupta et al., 2001).
Nixon and Ulmann (2006) in a study on EU countries within
1980-1995 showed health expenditures and physician numbers have significant
contribution to improvement in infant mortality and life expectancy (Nixon
and Ulmann, 2006). An important point about the effect of health expenditures
in our study, unlike to all of mentioned studies in this area, is that they
only investigated one aspect of health or different aspects of health separately,
thus, our study provides a different perspective about the effects health expenditures
on health status, especially in the developing countries.
According to the specific conditions of MENA region, some especial independent
variables i.e., immunization coverage, rate of access to safe
water and sanitation and particularly political stability as effective
factors on health status were modeled. The results confirmed that these variables
have significant effects on health outcomes particularly political stability
through its interactive effects on other effective variables. Wide aspects of
a society, particularly in developing countries, is affected by levels of stability
in political conditions such as income distribution, organizational structures,
values and how devoting of financial resources especially in health sector by
governors and policy makers. Starfield (2007) investigated
the pathways of the influence of equity in access to health facilities and fundamental
effects of political and policy contexts on health status (Starfield,
More importantly, present findings highlight the significance of political-structural
mechanisms and their interaction with national policies and institutions. According
to the results of the present research and previous studies on public health
challenges in MENA (Akala and El-Saharty, 2006; Pierre-Louis
et al., 2004), it can be concluded that one of the most important
challenges in public health systems in MENA region which is sensitive to country
circumstances yet reflects their regional context is distinction between
services and function that mainly steams from weak performance by governments
affecting by political instability (indoor and outdoor conflicts) and its negative
effect on organizational structures and policy makers. Trying to optimizing
in management and monitoring of services, directing health expenditures toward
cost-effective programs and reducing the gaps in health coverage can effectively
improve health systems functioning in this region. Wagstaff
and Claeson (2004) showed that good policies and institutions are important
determinants of the impact of government health expenditures on health outcomes.
In particular, as the quality of policies and institutions improves, the impact
of government health expenditures on maternal, infant and under-five mortality
and underweight children also increases and is statistically significant.
STUDY STRENGTHS AND LIMITATIONS
Although, many studies on national and cross-national levels on the evaluation of health care systems have been done. However, such researches in the developing countries are rare and to our knowledge, the present study is the first with special focus on MENA region.
Our study as the first model based in the scope of comparative studies related
to health systems efficiency on mortality indicators in MENA, using a suitable
modeling approach i.e., latent class regression, provided a comprehensive framework
towards the health systems evaluation, through combining the three health outcomes
(LEB, IMR and U5MR) and evaluation the effects of a wide range of medical and
non-medical effective factors on health status over the relatively wide study
The most important limitation of the present study was lack of a complete and reliable data set over the study period for MENA countries particularly for Djibouti, Afghanistan, Somalia and Yemen. Therefore, the study was restricted to the years: 1985, 1990, 1995, 2000 and 2005 in which over these years the more reliable data about the all considered countries had been provided, especially, by international organizations. Due to, most data over the considered years for Afghanistan and Djibouti was missing these countries were not included in the study.
1: Agha, S., 2000. The determinant of infant mortality in Pakistan. Social Sci. Med., 51: 199-208.
2: Akala, F.A. and S. El-Saharty, 2006. Public-health challenges in the Middle East and North Africa. Lancet, 367: 961-964.
3: Behm, H., 1991. An Analytical Framework. In: Child Mortality in Developing Countries, Behm, H. (Ed.). 2nd Edn., Oxford University Press, New York, pp: 7-20.
4: Behrman, J.R. and B.L. Wolfe, 1984. More evidence on nutrition demand: Income seems overrated and women's schooling underemphasized. J. Dev. Econ., 14: 105-128.
5: Burnside, C. and D. Dollar, 1998. Aid, the incentive regime and poverty reduction. Policy Research Working Paper 1937, World Bank, Development Research Group, Washington DC., http://papers.ssrn.com/sol3/papers.cfm?abstract_id=597236.
6: Caldwell, J.C., 1986. Routes to low mortality in poor countries. Popul. Dev. Rev., 12: 171-220.
Direct Link |
7: Casterline, J.B., E.C. Cooksey and A.F.E. Ismail, 1989. Household income and child survival in Egypt. Demography, 26: 15-35.
Direct Link |
8: Currie, J. and E. Moretti, 2003. Mother's education and the intergenerational transmission of human capital: Evidence from college openings. Q. J. Econ., 118: 1495-1532.
9: Cutler, D.M., A.S. Deaton and A. Lleras-Muney, 2006. The determinants of mortality. National Bureau of Economic Research, http://www.nber.org/aginghealth/spring06/w11963.html.
10: Deb, P. and P.K. Trivedi, 2002. The structure of demand for health care: Latent class versus two-part models. J. Health Econ., 21: 601-625.
11: Filmer, D. and L. Pritchett, 1997. Child mortality and public spending on health: How much does money matter? Policy Research Working Paper 1864. The World Bank.
12: Filmer, D. and L. Pritchett, 1999. The impact of public spending on health: Does money matter?. Soc. Sci. Med., 49: 1309-1323.
13: Fogel, R.W., 1994. Economic growth, population theory and physiology: The bearing of long-term processes on the making of economic policy. Am. Econ. Rev., 84: 369-395.
Direct Link |
14: Fuse, K. and E.M. Crenshaw, 2006. Gender imbalance in infant mortality: A cross-national study of social structure and female infanticide. Soc. Sci. Med., 62: 360-374.
15: Grubaugh, S.G. and R.E. Santerre, 1994. Comparing the performance of health care systems: An alternative approach. Southern Econ. J., 60: 1030-1042.
Direct Link |
16: Gupta, S., M. Verhoeven and E. Tiongson, 1999. Does higher government spending buy better results in education and health care?. IMF Working Paper No. 99/21. http://www.imf.org/external/pubs/cat/longres.cfm?sk=2892.0.
17: Gupta, S., M. Verhoeven and E. Tiongson, 2001. Public spending on health care and the poor. Working Paper No. 01/127, International Monetary Fund. Washington, DC. http://www.imf.org/external/pubs/cat/longres.cfm?sk=15316.0.
18: Heerink, N., 1994. Population Growth, Income Distribution, Economic Development and Population Economics. Springer-Verlag, Berlin.
19: Hill, K. and A.R. Pebley, 1989. Child mortality in the developing world. Population Dev. Rev., 15: 657-687.
Direct Link |
20: Hitiris, T. and J. Posnett, 1992. The determinants and effects of health expenditure in developed countries. J. Health Econ., 11: 173-181.
CrossRef | PubMed |
21: Jimenez-Martin, S., J.M. Labeaga and M. Martinez-Granado, 2002. Latent class versus two-part models in the demand for physician services across the European Union. Health Econ., 11: 301-321.
22: Kassouf, A.L. and B. Senauer, 1996. Direct and indirect effects of parental education on malnutrition among children in Brazil: A full income approach. Econ. Dev. Cult. Change, 44: 817-838.
Direct Link |
23: Kim, K. and P.M. Moody, 1992. More resources better health? Across-national perspective. Soc. Sci. Med., 34: 837-842.
24: Pierre-Louis, M., F.A. Akala and H.S. Karam, 2004. Public Health in the Middle East and North Africa: Meeting the Challenges of the 21st Century. World Bank Group, Washington, DC.
25: Macinko, J., F.C. Guanais and M.F.M. de Souza, 2006. Evidence based public health policy and practice: Evaluation of the impact of the Family Health Program on infant mortality in Brazil, 1990-2002. J. Epidemiol. Community Health, 60: 13-19.
26: McGuire, J.W., 2006. Basic health care provision and under-5 mortality: A Cross-National study of developing countries. World Dev., 34: 405-425.
27: Mosley, W.H. and L.C. Chen, 1984. An analytical framework for the study of child survival in developing countries. Popul. Dev. Rev., 10: 25-45.
CrossRef | Direct Link |
28: Musgrove, P., 1996. Public and private roles in health: Theory and financing patterns. World Bank Discussion Paper No. 339. The World Bank, Washington, DC.
29: Nixon, J. and P. Ulmann, 2006. The relationship between health care expenditure and health outcomes: Evidence and caveats for a causal link. Eur. J. Health Econ., 7: 7-18.
Direct Link |
30: Madise, N.J., Z. Matthews and B. Margetts, 1999. Heterogeneity of child nutritional status between households: A comparison of six Sub-Saharan African countries. Population Stud., 53: 331-343.
Direct Link |
31: O'Hara, D.J., 1980. Toward a model of the effects of education on health. The effects of education on health. World Bank Staff Working Paper, 405, World Bank, Washington, DC.
32: Or, Z., 2001. Exploring the effects of health care on mortality across OECD countries. OECD Labour Marketand Social Policy Occasional Papers, No. 46, OECD Publishing. http://dx.doi.org/10.1787/716472585704.
33: Or, Z., J. Wang and D. Jamison, 2005. International differences in the impact of doctors on health: A multilevel analysis of OECD countries. J. Health Econ., 24: 531-560.
34: Palloni, A. and K. Hill, 1997. The Effect of Economic Changes on Mortality by Age and Cause: Latin America, 1950-90. In: Demographic Responses to Economic Adjustment in Latin America, Tapinos, G., A. Mason and J. Bravo (Eds.). Oxford University Press, Oxford, UK.
35: Papageorgiou, C., A. Savvides and M. Zachariadis, 2004. International medical R and D spillovers. Departmental Working Papers 2004-03, Department of Economics, Louisiana State University, http://www.bus.lsu.edu/economics/papers/pap04_03.pdf.
36: Preston, S.H., 1980. Causes and Consequences of Mortality Decline in Less Developed Countries During the Twentieth Century. In: Population and Economic Change in Developing Countries, Easterlin, R.A. (Ed.). University of Chicago Press, Chicago.
37: Robst, J., 2001. A note on the relationship between medical care resources and mortality. Applied Econ. Lett., 8: 737-739.
38: Robst, J. and G.G. Graham, 1997. Access to health care and current health status: Do physicians matter?. Applied Econ. Lett., 4: 45-48.
39: Sastry, N., 1997. What explainsrural-urban differentials in child mortality in Brazil?. Soc. Sci. Med., 44: 989-1002.
40: Smith, P., 2002. Developing Composite Indicators for Assessing Health System Efficiency. Measuring Up: Improving the Performance of Health Systems in OECD Countries. OECD, Paris.
41: Starfield, B., 2007. Pathways of influence on equity in health. Soc. Sci. Med., 64: 1355-1362.
CrossRef | PubMed |
42: Valkonen, T., 1989. Adult Mortality and Level of Education: A Comparison of Six Countries. In: Health Inequalities in European Countries, Fox J. (Ed.). Gower and European Science Foundation, Aldershot, pp: 142-160.
43: Wagstaff, A. and M. Claeson, 2004. The Millennium Development Goals for Health: Rising to the Challenges. World Bank Publications, Washington, DC.
44: WHO, 2000. Health Systems: Improving Performance. World Health Organization, Geneva, Switzerland.
45: Zakir, M. and P.V. Wunnava, 1999. Factors affecting infant mortality rates: evidence from cross-sectional data. Applied Econ. Lett., 6: 271-273.