Socioeconomic determinants of infant mortality: a worldwide study of 152 low-, middle-, and high-income countries

BACKGROUND: To reach the Millennium Development Goals for health, influential international bodies advocate for more resources to be directed to the health sector, in particular medical treatment. Yet, health has many determinants beyond the health sector that are less evident than proximate predictors. AIM: To assess the relative importance of major socioeconomic determinants of population health, measured as infant mortality rate (IMR), at country level. METHODS: National-level data from 152 countries based on World Development Indicators 2003 were used for multivariate linear regression analyses of five socioeconomic predictors of IMR: public spending on health, GNI/capita, poverty rate, income equality (Gini index), and young female illiteracy rate. Analyses were performed on a global level and stratified for low-, middle-, and high-income countries. RESULTS: In order of importance, GNI/capita, young female illiteracy, and income equality predicted 92% of the variation in national IMR whereas public spending on health and poverty rate were non-significant determinants when adjusted for confounding. In low-income countries, female illiteracy was more important than GNI/capita. Income equality (Gini index) was an independent predictor of IMR in middle-income countries only. In high-income countries none of these predictors was significant. CONCLUSIONS: The relative importance of major health determinants varies between income levels, thus extrapolating health policies from high- to low-income countries is problematic. Since the size, per se, of public health spending does not independently predict health outcomes, functioning health systems are necessary to make health investments efficient. Potential health gains from improved female education and economic growth should be considered in low- and middle-income countries.     

Income inequality,illiteracy rate,and life expectancy in Brazil

OBJECTIVES: The link between income disparities and health has been studied mostly in developed nations. This study assesses the relationship between income disparities and life expectancy in Brazil and measures the impact of illiteracy rates on the association. METHODS: The units of analysis (n = 27) are all the Brazilian states and the federal capital. Simple and multiple linear regressions were performed to measure the association between income disparity, measured by the Gini coefficient, gross domestic product (GDP) per capita, and illiteracy rate. Data came from publicly available sources at the Brazilian Ministry of Health and the Brazilian Institute of Geography and Statistics. RESULTS: Income disparities and illiteracy rates were negatively associated with life expectancy in Brazil. GDP per capita was positively associated with life expectancy. The inclusion of illiteracy rates in the regression model removed the effect of income disparities. CONCLUSIONS: Illiteracy rate is strongly associated with life expectancy in Brazil. This finding is in accord with reports from the United States and has implications for health policy and planning for both developed and developing countries.     

Has the relation between income inequality and life expectancy disappeared? Evidence from Italy and top industrialised countries

OBJECTIVE: To investigate the relation between income inequality and life expectancy in Italy and across wealthy nations. DESIGN AND SETTING: Measure correlation between income inequality and life expectancy at birth within Italy and across the top 21 wealthy countries. Pearson correlation coefficients were calculated to study these relations. Multivariate linear regression was used to measure the association between income inequality and life expectancy at birth adjusting for per capita income, education, and/or per capita gross domestic product. DATA SOURCES: Data on the Gini coefficient (income inequality), life expectancy at birth, per capita income, and educational attainment for Italy came from the surveys on Italian household on income and wealth 1995-2000 and the National Institute of Statistics information system. Data for industrialised nations were taken from the United Nations Development Program's human development indicators database 2003. RESULTS: In Italy, income inequality (beta = -0.433; p<0.001) and educational attainment (beta = 0.306; p<0.001) were independently associated with life expectancy, but per capita income was not (beta = 0.121; p>0.05). In cross national analyses, income inequality had a strong negative correlation with life expectancy at birth (r = -0.864; p<0.001). CONCLUSIONS: In Italy, a country where health care and education are universally available, and with a strong social safety net, income inequality had an independent and more powerful effect on life expectancy at birth than did per capita income and educational attainment. Italy had a moderately high degree of income inequality and an average life expectancy compared with other wealthy countries. The cross national analyses showed that the relation between income inequality and population health has not disappeared.     

Trends in infant mortality inequalities in the Americas: 1955-1995

STUDY OBJECTIVE:s: To describe overall and income related trends in infant mortality inequalities in the Region of the Americas from 1955 to 1995. DESIGN: Infant mortality rates (IMRs) were computed and their trends assessed by ordinary least squares. Overall trends in IMR inequalities among countries were analysed by comparing 10 year period IMRs, Gini coefficients, and Lorenz curves. Income related trends in IMR inequalities were assessed using 10 year period IMR ratios between the highest and the lowest quintiles of the per capita gross national product (GNP) distributions (adjusted for purchasing power). SETTING: Aggregated country data were used for all countries with over 200 thousand inhabitants (33 geopolitical units). The 10 year period midpoint IMR estimates used for the 1955-1995 time series were those published by the United Nations in 1997. MAIN RESULTS: IMRs decreased from 90.34 to 31.31 per 1000 live births between 1955 and 1995 at an average of 15.3 every 10 years. In contrast, Lorenz curves and Gini coefficients were similar for the five 10 year periods. After grouping by adjusted GNP distribution, a similar decreasing trend of IMR was observed in all groups. The rate ratio between the group at the lowest quintile and that at the highest quintile ranged from 4 to 5. The analysis of variance for repeated observations showed that there is a significant reduction in the IMR (F=130.18; p<0.01), that trends did not differ significantly among groups (F=1.16; p=0.32), and that they were approximately linear (F=155.83; p<0.01). CONCLUSIONS: Despite a sizable reduction in the infant mortality, whether or not income related, levels of IMR inequality among countries have remained almost constant between 1955 and 1995 in the Region of the Americas. Further analysis and focused interventions are needed to tackle the challenges of reducing these persistent mortality inequalities.     

Societal characteristics and health in the former communist countries of Central and Eastern Europe and the former Soviet Union: a multilevel analysis

OBJECTIVES: To examine whether, in former communist countries that have undergone profound social and economic transformation, health status is associated with income inequality and other societal characteristics, and whether this represents something more than the association of health status with individual socioeconomic circumstances. DESIGN: Multilevel analysis of cross-sectional data. SETTING: 13 Countries from Central and Eastern Europe and the former Soviet Union. PARTICIPANTS: Population samples aged 18+ years (a total of 15 331 respondents). MEAN OUTCOME MEASURES: Poor self-rated health. RESULTS: There were marked differences among participating countries in rates of poor health (a greater than twofold difference between the countries with the highest and lowest rates of poor health), gross domestic product per capita adjusted for purchasing power parity (a greater than threefold difference), the Gini coefficient of income inequality (twofold difference), corruption index (twofold difference) and homicide rates (20-fold difference). Ecologically, the age- and sex-standardised prevalence of poor self-rated health correlated strongly with life expectancy at age 15 (r = -0.73). In multilevel analyses, societal (country-level) measures of income inequality were not associated with poor health. Corruption and gross domestic product per capita were associated with poor health after controlling for individuals' socioeconomic circumstances (education, household income, marital status and ownership of household items); the odds ratios were 1.15 (95% confidence interval 1.03 to 1.29) per 1 unit (on a 10-point scale) increase in the corruption index and 0.79 (95% confidence interval 0.68 to 0.93) per $5000 increase in gross domestic product per capita. The effects of gross domestic product and corruption were virtually identical in people whose household income was below and above the median. CONCLUSION: Societal measures of prosperity and corruption, but not income inequalities, were associated with health independently of individual-level socioeconomic characteristics. The finding that these effects were similar in persons with lower and higher income suggests that these factors do not operate exclusively through poverty.     

Democratisation and health after the fall of the Wall

OBJECTIVES: The fall of the Wall in 1989 can be seen as a natural experiment in the epidemiological sense to further examine the relation between democracy and health. DESIGN AND SETTING: Ecological study in the 23 post-communist countries, during the last decade of the 20th century, exploring the relations between the level of democratisation and health, taking into account as relevant confounders wealth and the level of inequality. MAIN RESULTS: A significant correlation (p<0.01) was found of the democratic deficit of the countries with the health indicators circa 2000, with values of Pearson's coefficient of -0.629 for life expectancy, 0.760 for infant mortality, and 0.555 for maternal mortality. These associations remain significant after adjustment by lineal regression for GNP per capita and the Gini coefficient, with R(2) values of 0.336 for life expectancy, 0.575 for infant mortality, and 0.529 for maternal mortality. CONCLUSIONS: These findings add pieces of evidence to the previously reported cross sectional association between democracy and health.     

A crossectional country group study of three health level outcomes and four potential causes.

A crossectional study of four country groups segmented by per capita income of the majority of the world's countries was made to evaluate the relationship between health level outcomes and potential causes which may impact on the health level outcomes. The health level outcomes consist of life expectancy at birth, infant mortality rate and child mortality rate. The potential causes consist of secondary school children per 100 in school age group, daily calory supply per capita, population per physician and population per nurse. For the two lower income country groups the two important determinants of life expectancy were daily calory supply per capita and secondary school children per 100 in school age group. For the upper middle income the country group the important positive determinant of life expectancy was population per nurse and for the upper income country group the important negative determinant of life expectancy was daily calory supply per capita. Infant and child mortality rates were associated with secondary school children per 100 in school age group and population per physician or population per nurse for the two lower income country groups. For the upper middle income country group population per nurse or population per physician was supplemented by daily calory supply per capita for both infant and child mortality. For the upper income country group only infant mortality had statistically significant determinants. They were daily calory supply per capita and secondary school children per 100 in school age groups.     

Income inequality and population health: correlation and causality

A large literature now exists on the cross-national correlation between income inequality and population health, but existing studies suffer from sparse data, poor operationalization of income inequality, and the use of low-power statistical models. This paper sets out to estimate the ecological correlation between income inequality and indicators of population health in a very broad panel of countries, to demonstrate that this relationship is largely non-artifactual, and to test whether this relationship might be causal. Gini coefficients of national income inequality in 1970 and 1995 are correlated with life expectancy, infant mortality rates, and murder rates, controlling for national income per capita. In cross-sectional analyses, inequality is significantly correlated with life expectancy, infant mortality, and (inconsistently) the murder rate. The health correlations are shown to be not primarily due to the "convexity effect" of the non-linear relationship between individual income and individual health, which seems to account for no more than one-third of the relationship between inequality and health, and likely much less. Change in inequality 1970-1995 is significantly related to change in life expectancy and infant mortality, suggesting a causal relationship, but these correlations are not robust with respect to sample or controls. It can be concluded that there is a strong, consistent, statistically significant, non-artifactual correlation between national income inequality and population health, but though there is some evidence that this relationship is causal, the relative stability of income inequality over time in most countries makes causality difficult to test.     

Correlation between national income, HIV/AIDS and political status and mortalities in African countries

OBJECTIVES: To investigate associations between mortalities in African countries and problems that emerged in Africa in the 1990s (reduction of national income, HIV/AIDS and political instability) by adjusting for the influences of development, sanitation and education. METHODS: We compiled country-level indicators of mortalities, national net income (the reduction of national income by the debt), infection rate of HIV/AIDS, political instability, demography, education, sanitation and infrastructure, from 1990 to 2000 of all African countries (n=53). To extract major factors from indicators of the latter four categories, we carried out principal component analysis. We used multiple regression analysis to examine the associations between mortality indicators and national net income per capita, infection rate of HIV/AIDS, and political instability by adjusting the influence of other possible mortality determinants. RESULTS: Mean of infant mortality per 1000 live births (IMR); maternal mortality per 100,000 live birth (MMR); adult female mortality per 1000 population (AMRF); adult male mortality per 1000 population (AMRM); and life expectancy at birth (LE) in 2000 were 83, 733, 381, 435, and 51, respectively. Three factors were identified as major influences on development: education, sanitation and infrastructure. National net income per capita showed independent negative associations with MMR and AMRF, and a positive association with LE. Infection rate of HIV/AIDS was independently positively associated with AMRM and AMRF, and negatively associated with LE in 2000. Political instability score was independently positively associated with MMR. CONCLUSIONS: National net income per capita, HIV/AIDS and political status were predictors of mortality indicators in African countries. This study provided evidence for supporting health policies that take economic and political stability into account.     

The evaluation of educational and economic effects on life expectation by linear regression analysis

Socioeconomic effects are supposed to have a fairly long time delay in regard to health development. Taking into consideration time lags, correlation and regression analysis are executed to explain the difference in health levels which exist among countries. The infant mortality rate, life expectation at birth, adult literacy rate, proportions of school enrollment and GNP per capita are taken as variables and their characteristics are discussed. In addition, countries are divided into three groups, i.e. low income, middle income, and industrialized countries, to show that the educational and economic effects on health development vary according to the economic level of the countries.     

State size as measured in terms of public spending and world health, 1990-2000

OBJECTIVE: To determine the relationship between state size (measured in terms of public spending) and public health indicators in a sample of countries representing all regions of the world and from 1990-2000. METHODS: An ecological study was performed using data on Central Government Spending (CGS) and per capita Gross National Product (GNP) obtained from the International Monetary Fund, and on life expectancy, maternal, and infant mortality, provided by the World Health Organization. A multiple linear regression model was fitted to estimate the effect of CGS on health, which also took into consideration per capita GNP and geographical region. RESULTS: CGS varied little over the study period, with convergence around an average of 28%, but within a relatively wide range (7.80-53.0%); the countries with the strongest economies (according to per capita GNP) had the highest levels of CGS. The influence of this factor was particularly relevant for the infant mortality rate (r = 0.40; beta = -1.327; EE = 0.237; t = -5.590; p < 0.001). Per capita GNP and geographic location were also associated with variations in health; health indicators tended to be worse for poorer countries in Africa and Asia. In the adjusted model, CGS was statistically significant with regard to infant and maternal mortality rates. CONCLUSION: The study suggests that state size (in terms of public spending) has an important influence upon health and particularly upon mortality. Although it is important to bear in mind the limitations of this study and the reduced time window used, these results should be taken into consideration in the current political and epidemiological debate.     

Economic development, political-economic system, and the physical quality of life.

This study compared capitalist and socialist countries in measures of the physical quality of life (PQL), taking into account the level of economic development. The World Bank was the principal source of statistical data for 123 countries (97 per cent of the world's population). PQL variables included: indicators of health, health services, and nutrition (infant mortality rate, child death rate, life expectancy, population per physician, population per nursing person, and daily per capita calorie supply); measures of education (adult literacy rate, enrollment in secondary education, and enrollment in higher education); and a composite PQL index. Capitalist countries fell across the entire range of economic development (measured by gross national product per capita), while the socialist countries appeared at the low-income, lower-middle-income, and upper-middle-income levels. All PQL measures improved as economic development increased. In 28 of 30 comparisons between countries at similar levels of economic development, socialist countries showed more favorable PQL outcomes.     

The importance of the political and the social in explaining mortality differentials among the countries of the OECD, 1950-1998.

This article analyzes (within the conceptual frame defined in the previous article) the impact of political variables such as time of government by political parties (social democratic, Christian democratic or conservative, liberal, and ex-dictatorial that have governed the OECD countries during the 1950-1998 period) and their electoral support on (1) redistributional policies in the labor market and in the welfare state; (2) the income inequalities measured by Theil and Gini indexes; and (3) health indicators, such as infant mortality and life expectancy. This analysis is carried out statistically by a bivariate and a multivariate analysis (a pooled cross-sectional study). Both analyses show that political variables play an important role in defining how public and social policies determine the levels of inequalities and affect the level of infant mortality. In general, political parties more committed to redistributional policies, such as social democratic parties, are the most successful in reducing inequalities and improving infant mortality. Less evidence exists, however, on effects on life expectancy. The article also quantifies statistically the relationship between the political and the policy variables and between these variables and the dependent variables--that is, the health indicators.     

Influence of income, income inequalities and social capital on the health of persons aged 65 and over in Spain in 2007

The aim of this study is to evaluate the influence of personal income [absolute income hypothesis (AIH)], income inequality and welfare [relative income hypothesis (RIH)], and social capital on the health of older people. Multi-level, cross-sectional logit models are calculated separately for women and men. The database employed was the Spanish Life Conditions Survey for 2007. The population consists of 6,259 persons aged over 65 years living in the 17 autonomous regions of Spain. The results confirm the AIH hypothesis: higher personal income is associated with better health. Education is also associated with better self-perceived health. The RIH hypothesis is partially confirmed due to the association between the Gini coefficient, regional per capita welfare and self-perceived health in older people, but only for women. Two different measures of social capital are used: the value of services of social capital and the percentage of people aged over 65 belonging to an association. Both factors are statistically associated with better self-perceived health in women. This study is the first to contrast the associations among income, income inequalities, social capital and the health of elders in Spain.     

我国人均卫生费用的主成分分析

目的:分析我国人均卫生费用的影响因素,去除线性回归中多变量相关问题带来的不利影响。方法:基于1990-2014年我国人均卫生费用,选取人均GDP、人均可支配收入、城镇化率、人口老龄化、婴儿死亡率、每千人口执业医师数和每千人口医疗卫生机构床位数作为解释变量,使用主成分分析模型研究解释变量如何影响我国人均卫生费用。结果:解释变量均显著影响人均卫生费用。其中,人均GDP、人均可支配收入、每千人口执业医师数和每千人口医疗卫生机构床位数正向影响人均卫生费用,婴儿死亡率、城镇化率和人口老龄化负向影响人均卫生费用。结论:我国人均卫生费用的影响因素并非是单一的。人均卫生费用的上涨是多个因素的共同结果。     

Trends in the association between average income, poverty and income inequality and life expectancy in Spain

In this paper, we study the relation between life expectancy and both average income and measures of income inequality in 1980 and 1990, using the 17 Spanish regions as units of analysis. Average income was measured as average total income per household. The indicators of income inequality used were three measures of relative poverty-the percentage of households with total income less than 25%, 40% and 50% of the average total household income-the Gini index and the Atkinson indices with parameters alpha=1, 1.5 and 2. Pearson and partial correlation coefficients were used to evaluate the association between average income and measures of income inequality and life expectancy. None of the correlation coefficients for the association between life expectancy and average household income was significant for men. The association between life expectancy and average household income in women, adjusted for any of the measures of income inequality, was significant in 1980, although this association decreased or disappeared in 1990 after adjusting for measures of poverty. In both men and women, the partial correlation coefficients between life expectancy and the measures of relative income adjusted for average income were positive in 1980 and negative in 1990, although none of them was significant. The results with regard to women confirm the hypothesis that life expectancy in the developed countries has become more dissociated from average income level and more associated with income inequality. The absence of a relation in men in 1990 may be due to the large impact of premature mortality from AIDS in regions with the highest average total income per household and/or smallest income inequality.     

Correlates of life expectancy in less developed countries

Analyses were performed to investigate several hypotheses concerning the multiple determinants of levels of life expectancy in developing countries in recent decades and some possible explanation for the observed variations in amount of gain in life expectancy from the 1950's to the 1970's. The findings were significant. For level of life expectancy the results of this present work conform by and large to results of other scholars in this area, although the present work is unique in that only developing countries were included. From the 1960's to the 1970's there has been a shift in the relative importance of economic indicators and general social indicators in favor of the social indicators. In the period 1960-65 some 70% of the variation in levels of life expectancy was associated with per capita income and literacy rates in a ratio of about three to two in favor of the economic variable. By 1970-75 the ratio has become six to one in favor of literacy. In addition, the multivariate model showed that the sanitation variables began to appear as significant correlates of levels of life expectancy in the more recent time period, playing a larger role than level of income per capita. Work pursued as part of a separate but concurrent project explored explicitly this three-way interaction between literacy, life expectancy and sanitation.     

Linkages between biodiversity loss and human health: a global indicator analysis

The association between health and biodiversity loss was explored by means of regression analysis on a global scale, with control for confounding by socio-economic developments. For this we selected indicators of human health (life expectancy, disability adjusted life expectancy, infant mortality rate and percentage low-birthweight babies), biodiversity (percentage threatened species, current forest as a percentage of original forest, percentage of land highly disturbed by man) and socio-economic development (health expenditure as percentage of GNP, percentage one-year olds immunized, illiteracy rate, GNP per capita and development grade) on a country level. After controlling for relevant socio-economic confounders, both current forest as a percentage of original forest and the percentage of land highly disturbed by human activities had no relationship with one of the health indicators. The logarithm of the percentage threatened species, showed a positive association with life expectancy and disability adjusted life expectancy. The present study was not able to provide any empirical proof of a negative association between loss of biodiversity and human health at the global scale. This does not mean, however, that no such relationship exists, because there may be several reasons for our findings, like possible non-linearity of the relationship, lack of suitable indicators, non-randomness in the sample of countries and the limitations of regression analysis in proving causality.     

期望寿命与婴儿死亡率的预测

[目的]预测中国婴儿死亡率与期望寿命。[方法]以人均国民收入(购买力平价)为自变量,同期各国婴儿死亡率为应变量,作曲线拟合,寻找最优预测模型;以婴儿死亡率为自变量,同期各国期望寿命为应变量,作线性回归。[结果]建立乘幂函数模型预测婴儿死亡率,决定系数为0.858,模型拟合效果较好。2007年中国婴儿死亡率预测值为20.86‰,实际报告值为19.00‰;建立线性回归模型预测期望寿命,调整决定系数为0.926,模型拟合效果较好。2007中国人口期望寿命预测值为72.36岁,报告值为74岁。预测2020年中国人口婴儿死亡率为8.46‰,期望寿命为75.25岁。[结论]2007年中国人均国民收入偏低,但是在降低婴儿死亡率和提高期望寿命方面优于同等经济实力国家的平均水平。择选与预测指标关系最为紧密的变量,保证足够大的样本含量,通过传统回归方法同样可以较好预测出婴儿死亡率和期望寿命。     

Income inequality, social capital and self-rated health and dental status in older Japanese

The erosion of social capital in more unequal societies is one mechanism for the association between income inequality and health. However, there are relatively few multi-level studies on the relation between income inequality, social capital and health outcomes. Existing studies have not used different types of health outcomes, such as dental status, a life-course measure of dental disease reflecting physical function in older adults, and self-rated health, which reflects current health status. The objective of this study was to assess whether individual and community social capital attenuated the associations between income inequality and two disparate health outcomes, self-rated health and dental status in Japan. Self-administered questionnaires were mailed to subjects in an ongoing Japanese prospective cohort study, the Aichi Gerontological Evaluation Study Project in 2003. Responses in Aichi, Japan, obtained from 5715 subjects and 3451 were included in the final analysis. The Gini coefficient was used as a measure of income inequality. Trust and volunteering were used as cognitive and structural individual-level social capital measures. Rates of subjects reporting mistrust and non-volunteering in each local district were used as cognitive and structural community-level social capital variables respectively. The covariates were sex, age, marital status, education, individual- and community-level equivalent income and smoking status. Dichotomized responses of self-rated health and number of remaining teeth were used as outcomes in multi-level logistic regression models. Income inequality was significantly associated with poor dental status and marginally significantly associated with poor self-rated health. Community-level structural social capital attenuated the covariate-adjusted odds ratio of income inequality for self-rated health by 16% whereas the association between income inequality and dental status was not substantially changed by any social capital variables. Social capital partially accounted for the association between income inequality and self-rated health but did not affect the strong association of income inequality and dental status.