Moving from Description to Action: Challenges in Researching Socio-Economic Inequalities in Health
Tugwell, Peter, Kristjansson, Betsy, Canadian Journal of Public Health
The Romanow report emphasized that Canada is committed to the goal of equity in health care and health outcomes.1 Unfortunately, we are a long way from meeting this objective. As in most other countries, health in Canada is unevenly and, to some extent, unfairly distributed according to socio-economic position. Health and longevity are highest for the richest and most powerful, and decrease with decreasing income.2-5 These social gradients in health, or health inequalities, are pervasive and persistent,6 and are seen in most diseases, injuries, and health behaviours.7 Health inequities are unfair and avoidable inequalities.8-10 Health inequities represent needless human suffering and have significant consequences for the economy and for social order and justice.4,11 It is time to take action and, indeed, social gradients in health are beginning to get the attention they deserve from policy-makers.
In this issue of the CJPH, Finkelstein demonstrates that those who research health inequalities face important challenges.12 We briefly describe some challenges associated with measuring health inequalities, situate this in the broader research agenda and identify the need for more accurate measures.
Finkelstein's paper highlights issues of definition and measurement of health inequalities and inequities. Accurate and consistent definition and measurement are crucial: they affect our results, interpretation, and, ultimately, our actions. Health inequalities can be objectively measured, but health inequities require normative judgements as to what constitutes 'unfairness' in the distribution of health and in the resources necessary for health.8,13 Our ultimate interest must lie in identifying and addressing health inequities, but measurement of health inequality is a prerequisite for this. Health inequalities require three components for calculation: a measure of health status, a measure of social position, and a statistical method for summarizing the magnitude of the health differences between people in different social positions.14
Measuring social position is challenging. Typically, indicators of socio-economic position (e.g., income, education, or occupational social class) are used. However, depending on the setting, other indicators of social stratification are relevant.13 The Global Equity Gauge Alliance has adopted the term PROGRESS15; this stands for Place of residence, Race/ethnicity, Occupation, Gender, Religion, Education, Socio-economic Status (income or composite measures), and Social Capital - to describe relevant indicators.16 We feel that measurement of health inequalities should always be based on indicators of socio-economic position (SEP); in Canada, rural/urban residence and ethnicity are also highly relevant. It is particularly important to consider the interrelationship between socio-economic position and other indicators of social stratification such as gender or ethnicity.13 For example, in the Alameda County Study, the shape of socio-economic gradients in morbidity differed substantially between men and women. There was an inverse relationship between income and health for both sexes, but this relationship was usually linear for women and nonlinear for men; low income men showed disproportionately high morbidity.17
Finkelstein's article emphasizes the difficulty researchers sometimes encounter in finding valid data on socio-economic position. Although Statistics Canada routinely includes measures of individual (including household) SRP in national health surveys, many administrative databases lack this information. Moreover, some public organizations that conduct surveys or evaluation studies are reluctant to ask about SEP due to fears about invasion of privacy. Many researchers are therefore forced to use measures of neighbourhood income or SES as proxies for individual level measures. But this may not always be an ideal solution. In Ontario, Finkelstein found that neighbourhoods were quite heterogeneous in terms of household income and that use of median Enumeration Area (EA) income as a proxy for household income resulted in high misclassification of household income. …