Socioeconomic inequalities in health: Measurement, computation, and statistical inference
TL;DR: In this article, the relationship between two widely used indices of health inequality and explain why these are superior to others indices used in the literature is explained and the role that demographic standardization plays in the analysis of socioeconomic inequalities in health.
About: This article is published in Journal of Econometrics.The article was published on 1997-03-01 and is currently open access. It has received 1250 citations till now. The article focuses on the topics: Health equity & Statistical inference.
Summary (1 min read)
Jump to: [I. Introduction] – [3. Demographic factors and avoidable inequality] – [5. Empirical illustrations] and [6. Conclusions]
I. Introduction
- This paper clarifies the relationship between two widely used indices of health inequality, namely the relative index of inequality (Rll) and the concentration index (CI).
- It explains why these are superior to other indices used in the empirical literature.
- Since the indices of health inequality are generally estimated from sample observations, it is useful to be able to test whether any observed differences in their values are statistically significant.
- This paper develops accurate distribution-free asymptotic estimators of the standard errors of both the R|I and CI.
- These sampling distributions are derived by applying Hoeffding's (1948) theorem on order statistics.
3. Demographic factors and avoidable inequality
- O which is negative if there are avoidable inequalities favouring the more (less) advantaged members of society.
- C* can be computed straightforwardly using Eq. ( 2) but replacing the actual illness score with the indirectly standardized score.
5. Empirical illustrations
- No significant avoidable inequalities in the presence of chronic illness but that there are significant avoidable inequalities in self-assessed health.
- These are the same conclusions that were reached with the grouped data in Section 4, though the t-values for chronic illness are a good deal smaller in the case of I* and I ÷.
6. Conclusions
- If the authors assume that the groups are homogeneous (which is the case when individual observations are available), tr~ will be zero in each group.
- The formulae of var(~) and var(fl) will simplify considerably.
Did you find this useful? Give us your feedback
Citations
More filters
Posted Content•
TL;DR: In this paper, the authors provide a step-by-step practical guide to the measurement of a variety of aspects of health equity, including gaps in health outcomes between the poor and the better-off in specific countries or in the developing world as a whole.
Abstract: This book shows how to implement a variety of analytic tools that allow health equity - along different dimensions and in different spheres - to be quantified. Questions that the techniques can help provide answers for include the following: Have gaps in health outcomes between the poor and the better-off grown in specific countries or in the developing world as a whole? Are they larger in one country than in another? Are health sector subsidies more equally distributed in some countries than in others? Is health care utilization equitably distributed in the sense that people in equal need receive similar amounts of health care irrespective of their income? Are health care payments more progressive in one health care financing system than in another? What are catastrophic payments? How can they be measured? How far do health care payments impoverish households? This volume has a simple aim: to provide researchers and analysts with a step-by-step practical guide to the measurement of a variety of aspects of health equity. Each chapter includes worked examples and computer code. The authors hope that these guides, and the easy-to-implement computer routines contained in them, will stimulate yet more analysis in the field of health equity, especially in developing countries. They hope this, in turn, will lead to more comprehensive monitoring of trends in health equity, a better understanding of the causes of these inequities, more extensive evaluation of the impacts of development programs on health equity, and more effective policies and programs to reduce inequities in the health sector.
1,301 citations
TL;DR: This paper aims to clarify the concepts of health disparities/inequalities (used interchangeably here) and health equity, focusing on the implications of different definitions for measurement and hence for accountability.
Abstract: There is little consensus about the meaning of the terms "health disparities," "health inequalities," or "health equity." The definitions can have important practical consequences, determining the measurements that are monitored by governments and international agencies and the activities that will be supported by resources earmarked to address health disparities/inequalities or health equity. This paper aims to clarify the concepts of health disparities/inequalities (used interchangeably here) and health equity, focusing on the implications of different definitions for measurement and hence for accountability. Health disparities/inequalities do not refer to all differences in health. A health disparity/inequality is a particular type of difference in health (or in the most important influences on health that could potentially be shaped by policies); it is a difference in which disadvantaged social groups-such as the poor, racial/ethnic minorities, women, or other groups who have persistently experienced social disadvantage or discrimination-systematically experience worse health or greater health risks than more advantaged social groups. ("Social advantage" refers to one's relative position in a social hierarchy determined by wealth, power, and/or prestige.) Health disparities/inequalities include differences between the most advantaged group in a given category-e.g., the wealthiest, the most powerful racial/ethnic group-and all others, not only between the best- and worst-off groups. Pursuing health equity means pursuing the elimination of such health disparities/inequalities.
1,214 citations
TL;DR: Two threshold approaches to measuring the fairness of health care payments are presented, one requiring that payments do not exceed a pre-specified proportion of pre-payment income, the other that they do not drive households into poverty, and the incidence and intensity of 'catastrophe' payments were reduced and became less concentrated among the poor.
Abstract: This paper presents and compares two threshold approaches to measuring the fairness of health care payments, one requiring that payments do not exceed a pre-specified proportion of pre-payment income, the other that they do not drive households into poverty. We develop indices for 'catastrophe' that capture the intensity of catastrophe as well as its incidence and also allow the analyst to capture the degree to which catastrophic payments occur disproportionately among poor households. Measures of poverty impact capturing both intensity and incidence are also developed. The arguments and methods are empirically illustrated with data on out-of-pocket payments from Vietnam in 1993 and 1998. This is not an uninteresting application given that 80% of health spending in that country was paid out-of-pocket in 1998. We find that the incidence and intensity of 'catastrophic' payments - both in terms of pre-payment income as well as ability to pay - were reduced between 1993 and 1998, and that both incidence and intensity of 'catastrophe' became less concentrated among the poor. We also find that the incidence and intensity of the poverty impact of out-of-pocket payments diminished over the period in question. Finally, we find that the poverty impact of out-of-pocket payments is primarily due to poor people becoming even poorer rather than the non-poor being made poor, and that it was not expenses associated with inpatient care that increased poverty but rather non-hospital expenditures.
979 citations
Additional excerpts
...[40]:...
[...]
Posted Content•
TL;DR: Wagstaff et al. as mentioned in this paper proposed a method for decomposing inequalities in the health sector into their causes, by coupling the concentration index with a regression framework, and showed how changes in inequality over time, and differences across countries, can be decomposed into the following: - Changes due to changing inequalities in determinants of the variable of interest. - Changes in the means of the determinants.
Abstract: A method for decomposing inequalities in the health sector into their causes is developed and applied to data on child malnutrition in Vietnam. Wagstaff, van Doorslaer, and Watanabe propose a method for decomposing inequalities in the health sector into their causes, by coupling the concentration index with a regression framework. They also show how changes in inequality over time, and differences across countries, can be decomposed into the following: - Changes due to changing inequalities in the determinants of the variable of interest. - Changes in the means of the determinants. - Changes in the effects of the determinants on the variable of interest. The authors illustrate the method using data on child malnutrition in Vietnam. They find that inequalities in height-for-age in 1993 and 1998 are accounted for largely by inequalities in household consumption and by unobserved influences at the commune level. And they find that an increase in such inequalities is accounted for largely by changes in these two influences. In the case of household consumption, rising inequalities play a part, but more important have been the inequality - increasing effects of rising average consumption and the increased protective effect of consumption on nutritional status. In the case of unobserved commune-level influences, rising inequality and general improvements seem to have been roughly equally important in accounting for rising inequality in malnutrition. This paper - a joint product of Public Services for Human Development, Development Research Group, and the Development Data Group - is part of a larger effort in the Bank to investigate the links between health and poverty. The authors may be contacted at awagstaff@worldbank.org, vandoorslaer@econ.bmg.eur.nl., or nwatanabe@worldbank.org.
853 citations
TL;DR: Equity in physician utilization favouring patients who are better off in about half of the OECD countries studied is found, and pro-rich inequity in doctor use is highest in the United States and Mexico, followed by Finland, Portugal and Sweden.
Abstract: Background: Most of the member countries of the Organization for Economic Cooperation and Development (OECD) aim to ensure equitable access to health care. This is often interpreted as requiring that care be available on the basis of need and not willingness or ability to pay. We sought to examine equity in physician utilization in 21 OECD countries for the year 2000. Methods: Using data from national surveys or from the European Community Household Panel, we extracted the number of visits to a general practitioner or medical specialist over the previous 12 months. Visits were standardized for need differences using age, sex and reported health levels as proxies. We measured inequity in doctor utilization by income using concentration indices of the need-standardized use. Results: We found inequity in physician utilization favouring patients who are better off in about half of the OECD countries studied. The degree of pro-rich inequity in doctor use is highest in the United States and Mexico, followed by Finland, Portugal and Sweden. In most countries, we found no evidence of inequity in the distribution of general practitioner visits across income groups, and where it does occur, it often indicates a pro-poor distribution. However, in all countries for which data are available, after controlling for need differences, people with higher incomes are significantly more likely to see a specialist than people with lower incomes and, in most countries, also more frequently. Pro-rich inequity is especially large in Portugal, Finland and Ireland. Interpretation: Although in most OECD countries general practitioner care is distributed fairly equally and is often even pro-poor, the very pro-rich distribution of specialist care tends to make total doctor utilization somewhat pro-rich. This phenomenon appears to be universal, but it is reinforced when private insurance or private care options are offered.
837 citations
Cites methods from "Socioeconomic inequalities in healt..."
...Robust estimates of the concentration index and HI index and its standard error can easily be obtained by running a convenient (weighted least squares) regression of a transformation of the variable on relative rank in the income distribution.(9) When the HI index equals zero, it indicates horizontal equity: people in equal need (but at different incomes) are treated equally....
[...]
References
More filters
Book•
01 Jan 1965TL;DR: Algebra of Vectors and Matrices, Probability Theory, Tools and Techniques, and Continuous Probability Models.
Abstract: Algebra of Vectors and Matrices. Probability Theory, Tools and Techniques. Continuous Probability Models. The Theory of Least Squares and Analysis of Variance. Criteria and Methods of Estimation. Large Sample Theory and Methods. Theory of Statistical Inference. Multivariate Analysis. Publications of the Author. Author Index. Subject Index.
8,300 citations
TL;DR: In this article, the authors considered the problem of estimating a U-statistic of the population characteristic of a regular functional function, where the sum ∑″ is extended over all permutations (α 1, α m ) of different integers, 1 α≤ (αi≤ n, n).
Abstract: Let X 1 …, X n be n independent random vectors, X v = , and Φ(x 1 …, x m ) a function of m(≤n) vectors . A statistic of the form , where the sum ∑″ is extended over all permutations (α1 …, α m ) of different integers, 1 α≤ (αi≤ n, is called a U-statistic. If X 1, …, X n have the same (cumulative) distribution function (d.f.) F(x), U is an unbiased estimate of the population characteristic θ(F) = f … f Φ(x 1,…, x m ) dF(x 1) … dF(x m ). θ(F) is called a regular functional of the d.f. F(x). Certain optimal properties of U-statistics as unbiased estimates of regular functionals have been established by Halmos [9] (cf. Section 4)
2,439 citations
TL;DR: Causal inference in statistics: An overview Linear Statistical Inference And Its Bayesian inference Wikipedia Springer Series in Statistics Stanford University Statistical Modeling, Causal Inference, and Social Science.
Abstract: Introduction to Statistical Inference Duke University Statistical Inference an overview | ScienceDirect Topics Bayesian statistics Amazon.com: The Elements of Statistical Learning: Data ... Causal Inference for Statistics, Social, and Biomedical ... Statistical Science Duke University STATISTICS BIOSTATISTICS Amazon.com: Statistical Inference (9780534243128): George ... CRAN Task View: Bayesian Inference Trevor Hastie Publications Stanford University Statistical model Wikipedia Psychology | UCLA Graduate Programs Linear Regression and Modeling | Coursera Supervised learning: predicting an output variable from ... Causal inference in statistics: An overview Linear Statistical Inference And Its Bayesian inference Wikipedia Springer Series in Statistics Stanford University Statistical Modeling, Causal Inference, and Social Science
1,702 citations
TL;DR: It is suggested that only two methods--the slope index of inequality and the concentration index--are likely to present an accurate picture of socioeconomic inequalities in health.
1,597 citations
TL;DR: Despite an overall decline in death rates in the United States since 1960, poor and poorly educated people still die at higher rates than those with higher incomes or better educations, and this disparity increased between 1960 and 1986.
Abstract: Background There is an inverse relation between socioeconomic status and mortality. Over the past several decades death rates in the United States have declined, but it is unclear whether all socioeconomic groups have benefited equally. Methods Using records from the 1986 National Mortality Followback Survey (n = 13,491) and the 1986 National Health Interview Survey (n = 30,725), we replicated the analysis by Kitagawa and Hauser of differential mortality in 1960. We calculated direct standardized mortality rates and indirect standardized mortality ratios for persons 25 to 64 years of age according to race, sex, income, and family status. Results The inverse relation between mortality and socioeconomic status persisted in 1986 and was stronger than in 1960. The disparity in mortality rates according to income and education increased for men and women, whites and blacks, and family members and unrelated persons. Over the 26-year period, the inequalities according to educational level increased for whites an...
1,517 citations