This section presents indicators and methods of measuring health equity.
Turrell G. and C. Mathers (2001). Socioeconomic inequalities in all-cause and specific-cause mortality in Australia: 1985–1987 and 1995–1997. International Journal of Epidemiology 30(2):231–9.
This study assesses trends in mortality inequality based on socioeconomic status in Australia for men and women aged 0–14, 15–24 and 25–64 years over the period 1985–1987 to 1995–1997 [13]. Socioeconomic status (SES) was operationalized using the Index of Relative Socioeconomic Disadvantage, an area-based measure developed by the Australian Bureau of Statistics. Mortality differentials were examined using age-standardized rates, and mortality inequality was assessed using rate ratios, gini coefficients, and excess mortality measures. For each period and for each sex/age subgroup, death rates were highest in the most disadvantaged areas, but the extent and nature of socioeconomic mortality inequality differed by sex and age group. If it were possible to reduce death rates among the SES areas to a level equivalent to that of the least disadvantaged area, premature all-cause mortality for men in each age group would be lower by 22%, 28% and 26% respectively, and for women, by 35%, 70% and 56%. A mixed pattern appears when examining the change in mortality inequality over the ten-year period. Among women, there was a decline in all-cause mortality inequality for each age group over the ten-year period, while cause-specific mortality inequalities increased for SIDS and traffic accidents. Among men, all-cause mortality inequalities increased over the ten-year period for age groups 0–14 and 15–24, but they decreased for those aged 25–64. For men, cause-specific mortality inequality increased for every condition except perinatal conditions and for drug dependence in those aged 10–14 and 15–24 years. The authors conclude, 1) "the mortality burden in the Australian population attributable to socioeconomic inequality is large and has profound and far-reaching implications in terms of unnecessary loss of life, loss of potentially economically productive members of society, and increased costs for the health care system"; but 2) the "simultaneous occurrence of widening, narrowing, and unchanging inequalities [over time]...is difficult to explain...on the basis of [a single] broad-ranging societal-level explanation."
Wolfson, M. and G. Rowe (2001). On measuring inequalities in health. Bulletin of the World Health Organization 79 (6): 553–560.
The authors present an alternative approach to measuring health inequalities, in contrast to the methods used in the World Health Report 2000[12]. Health inequalities can be conceived of in two different ways. The univariate or unconditional approach looks only at the health of individuals and views inequalities in health as the dispersion of health status within a population. Whereas the bivariate or conditional approach seeks to establish the distribution of health within a population, but conditional on another factor – whether those with low income also have poorer health, for example. The authors criticize the World Health Report approach for advocating univariate approaches, because they do not indicate the causes and social patterning of variations in health. The authors claim an even more significant weakness of the WHO approach lies in its proposed data collection strategy based on small area data. Several conceptual and methodological shortcomings limit the use of small area data, including non-random migration to or from the area, the small number of events such as deaths likely to be observed, and the likelihood that any specific geographical area may be associated with unique social, economic, or political conditions rendering it non-representative of the general population. The authors propose the use of longitudinal cohort-based data combined with micro-simulation-based life table analysis as a more fruitful analytic strategy.
Daniels, N., Bryant, J., Castano, R.A., Dantes, O.G., Khan, K.S., and S. Pannarunothai (2000). Benchmarks of fairness for health care reform: a policy tool for developing countries. Bulletin of the World Health Organization 78(6):740–50.
Teams of collaborators from Colombia, Mexico, Pakistan, and Thailand have adapted a policy tool originally developed for evaluating health insurance reforms in the United States into "benchmarks of fairness" for assessing health system reform proposals in developing countries [47]. The benchmarks include: intersectoral public health services and systems; financial and nonfinancial barriers to access to services; comprehensiveness and equity of benefits; equitable financing; efficacy, efficiency and quality of care; administrative efficiency; democratic accountability and empowerment; and patient and provider autonomy. Potential reforms are then evaluated by scoring (using either a "plus" or "minus" sign or a scale of -5 to 5, with zero representing the status quo) according to the degree to which they improve each criterion. The objective is "to promote discussion about fairness across the disciplinary divisions that keep policy analysts and the public from understanding how trade-offs between different effects of reforms can affect the overall fairness of the reform." The approach makes no effort to develop a uniform fairness scale across health systems, but could be used as a complement to assessments that do rank different countries according to objective standards of fairness. A striking feature of the criteria and rating process is the wide agreement on the benchmarks among the collaborating sites, despite their large historical, political and cultural differences.
See also Caplan, Light, and Daniels (1999) [48], where the authors discuss the benchmark approach for health equity in industrialized countries.
Gissler, M., Keskimaki, I., Teperi, J., Jarvelin, M., and E. Hemminki (2000). Regional equity in childhood health – register-based follow-up of the Finnish 1987 birth cohort. Health & Place 6(4):329–36.
In Finland, most surveillance of equity has been performed on adults [49]. This study investigated the extent to which regional health differences among Finnish children could be measured by using population-based longitudinal administrative register data. All children born in 1987 were included in the study (N = 59,546) and followed-up until the age of seven. Outcome measures included mortality, morbidity, and use of health services. Statistically significant regional variation was found for all health indicators but diabetes. Significant variation in use of health services was also found among all regions. Only in the case of mortality could variations be explained by confounders such as maternal age and social class. The authors emphasize that all of the variations observed do not necessarily imply inequity, since variation in genetic predisposition to disease, and greater use of services by those with greater need, would not be considered inequitable. They conclude that administrative registers offer a relatively inexpensive and quick means to monitor health equity, but that further work should go into developing more sensitive indicators of childhood health and service utilization.
van Doorslaer, E., Wagstaff, A., van der Burg, H., Christiansen, T., De Graeve, D., Duchesne, I., Gerdtham, U.G., Gerfin, M., Geurts, J., Gross, L., Hakkinen, U., John, J., Klavus, J., Leu, R.E., Nolan, B., O'Donnell, O., Propper, C., Puffer, F., Schellhorn, M., Sundberg, G., and O. Winkelhake (2000). Equity in the delivery of health care in Europe and the US. Journal of Health Economics 19(5):553–83.
This paper presents a comparison of horizontal equity in health care utilization in 10 European countries and the US [20]. It extends previous work by using more recent data from a larger set of countries, uses new methods and presents disaggregated results by various types of care. Horizontal equity is defined as "equal treatment for equal need". Health is measured by self-report and chronic ill-health. A concentration index measuring horizontal equity is constructed by "comparing each income group's share of need [for medical care] with its share of medical care [obtained]." Need is defined as "health care utilization that an individual on average is expected to receive given her age, gender, and various measures of self-reported health." In all countries, the lower-income groups are more intensive users of the general practitioner (GP) and the hospital. But after adjusting for increased medical need among the lower income groups, the authors find no overall indication of inequity. However, aggregate utilization masks important differences in the various components of medical care. In most of the countries studied, pro-rich inequity exists for physician contacts, because the rich have a higher than expected rate of use of specialist services compared to their health needs. The distribution of GP care across income groups is close to what is expected, although two countries (Belgium and Ireland) show greater than expected use of GPs by the poor – possibly because these countries exempt the poor from co-payments for GP visits. Hospital utilization is higher than expected among the poor, but due to imprecision in the estimates, this finding does not appear to be robust. The authors find no single health system feature (except that of co-payments for GP visits already discussed) to explain variations in equity among different health systems. They conclude, "in the late 1980s and early 1990s, the health care systems of [the countries studied] appeared to perform reasonably well on the horizontal equity criterion as applied in our methods."
Waters, H. R. (2000). Measuring equity in access to health care. Social Science and Medicine 51(4): 599–612.
The author proposes several methods to measure equity in access to health services and applies these measures in his analysis of the Ecuadorian General Health Insurance (GHI) program [21]. Equity in access measures include: 1) two egalitarian-based indicators for measuring equity in access to health care – a concentration coefficient derived from the Gini coefficient, and the Atkinson distributional measure; and 2) a weighted Utilitarian social welfare function to measure overall levels of access." The author discusses the derivation, calculation and interpretation of each equity measure in detail. The study found that Ecuador's "GHI program increases overall access to health care, but has a negative impact on equity in the distribution of health services". Potential policy options such as "expanding eligibility to the self-employed makes the benefit more equitably distributed (but still inequitable), and increases overall social welfare considerably. Expanding eligibility to the dependents of the insured person has similar effects."
Kinman, E. L. (1999). Evaluating health service equity at a primary care clinic in Chilimarca, Bolivia. Social Science and Medicine 49 (5): 663–78.
This paper links equity with a temporal and spatial analysis of clinic users, supplemented by a community survey [33]. Utilization of the primary care clinic in Chilimarca, Bolivia varied considerably during the first 25 months of operation. Spatially, utilization shifted away from the targeted service area. Within the targeted service area, usage was concentrated in a few blocks of the community and generally diminished with increasing distance from the clinic. The community survey revealed that place of origin, length of residence, and language spoken at home differentiated clinic users from non-users. Failure to include the spatial dimension of utilization would lead to different access and equity conclusions if data had not been decomposed by area. For example, over the period of the study, patients from the core catchment area declined by as much as 90 percent, to be replaced by clients from other areas. This resulted in changes in the average patient socio-demographic characteristics. The author concludes, "spatial analysis of output measures is imperfect and does not necessarily deal with all of the access issues related to acceptability. They do, however, begin to isolate areas of a defined geographic area where further investigation would assist in ascertaining, and subsequently addressing, potential problems related to equal access."
Lindholm, L., M. Rosen, and M. Emmelin (1998). How many lives is equity worth? A proposal for equity adjusted years of life saved. Journal of Epidemiology and Community Health 52(12): 808–11.
The objective of the article was to present a formula for equity adjusted years of life saved (EYLS) [6]. Swedish politicians responsible for health care in the county councils were given a scenario describing a trade-off between a health maximization program and one that is less efficient, but eliminates all social inequalities. The principle of health maximization was rejected. Under certain conditions, the Swedish politicians were prepared to sacrifice 15 out of 100 preventable deaths to achieve equity. Based on the results, a formula for EYLS was developed. Before it can be widely applied, the formula must be revised and adjusted to each country's specific conditions and values. The authors suggest that such formulas could be used to incorporate explicit considerations of equity into cost effectiveness analyses.
See also: Lindholm, Rosen, and Emmelin (1996) [50]. The authors find that at least two thirds of the Swedish politicians interviewed were prepared to accept lower growth in per capita health improvements in exchange for increased equity.
Kunst, A.E., Groenhof, F. and J.P. Mackenbach (1998). Mortality by occupational class among men 30–64 years in 11 European countries. Social Science and Medicine 46(11):1459–76.
The authors present evidence that within-country mortality differences between social classes are not necessarily smaller in European countries with more egalitarian socio-economic policies than in those with less egalitarian policies [14]. The authors compared eleven countries with respect to the magnitude of mortality differences by occupational class and paid particular attention to problems with the reliability and comparability of data available for different countries. Data problems were found to have the potential to substantially bias inequality estimates – especially those for Ireland, Spain and Portugal. In particular, problems in comparability of definitions of social class schemes, exclusion of the economically inactive men from the data sets, and discrepancies between social class definitions used on death certificates and census surveys may contribute to errors in measuring health inequalities. These differences in measurement may bias inequality estimates by up to 2 percent in England to 38 percent in Spain. When national mortality levels were considered, relatively large differences were observed for Finland and Ireland. The researchers found that the pattern of mortality differences varies from country to country and by age group, with the disparities being larger in northern countries than in southern ones (i.e. Italy, Spain, and Switzerland).
Manor, O., Matthews, S., and C. Power (1997). Comparing measures of health inequality. Social Science and Medicine 45:761–71.
The authors compared several methods of measuring social inequalities in health within different socioeconomic groups in Britain [15]. Health equity measures included 1) the slope or beta weight in multiple regression; 2) odds ratios; and 3) Agresti's alpha – an associational measure particularly useful for assessing health inequality when the health outcome variable is dichotomous. Each of these methods was compared using data from the British birth cohort. Inequalities in self-rated health, limited long-standing illness, psychological health, respiratory symptoms, asthma and obesity were calculated based on one of two measures of social position: class at birth and educational attainment. Results indicated that the magnitude of health inequalities did not differ significantly based on the type of health inequality measure used. However, the magnitude of health inequalities between groups did differ when such groups were constructed using different measures of social position; greater inequalities in health were detected between socioeconomic groups when such groups were defined by level of educational attainment rather than by social class at birth. Thus, how social class is specified makes a difference in drawing inferences about the magnitude of inequalities.
Wagstaff, A. (1991). QALYs and the equity-efficiency trade-off. Journal of Health Economics 10(1): 21–41.
As the volume of research on quality-adjusted life years (QALYs) has increased, concern has begun to be expressed about the equity aspects of resource allocation decisions based on the results of this research [51]. This paper suggests that a common theme running through the criticisms of the QALY approach is a concern about inequality. The paper describes methods for incorporating concerns about equity and the distribution of the burdens of disease into resource allocation decisions.
Musgrove, P. (1986). Measurement of equity in health. World Health Statistics Quarterly 39(4): 325–35.
This article discusses several approaches to measuring equity in resource distribution using data from Peru to illustrate each technique [22]. Equity is defined as "equal treatment for all of the population" and an equitable health care system is one that assures "probabilities [of access, given health need] will be equal across population groups for a given set of health problems". The author emphasizes that equity is "too complex a concept to be reduced to a single indicator" and proceeds to demonstrate techniques for measuring inequity in: 1) the distribution of health care resources such as physicians and hospital beds per capita within different geographic regions; 2) probabilities of treatment given medical need – which is sensitive to differences in type of illness studied, age group examined, and type of treatment investigated; 3) financial measures such as differences in expenditures adjusted for health need, or as a proportion of a household's total budget; and 4) indices such as the Gini coefficient for health care expenditures and availability of medical care. The author concludes that because assessments of equity (as opposed to inequality) require judgments about what is to be considered unfair, summary indicators of overall heath system inequity that do not capture the many ways in which inequity can be manifested (even within the same health system) are unlikely to inform interventions geared towards the improvement of inequities in health.
For further evidence and approaches to measuring health inequities, especially in developing countries, see the series of over 40 "Country reports on health, nutrition, population and poverty" produced by the World Bank and available on-line at http://www.worldbank.org/poverty/health/data/index.htm. For analyses of these data, see, Gwatkin and Guillot (2000) [52] and Wagstaff (2000) [53].