Income-related health inequality in Canada

https://doi.org/10.1016/S0277-9536(99)00319-6Get rights and content

Abstract

This study uses data from the 1994 National Population Health Survey and applies the methods developed by Wagstaff and van Doorslaer (1994, measuring inequalities in health in the presence of multiple-category morbidity indicators. Health Economics 3, 281–291) to measure the degree of income-related inequality in self-reported health in Canada by means of concentration indices. It finds that significant inequalities in self-reported ill-health exist and favour the higher income groups — the higher the level of income, the better the level of self-assessed health. The analysis also indicates that lower income individuals are somewhat more likely to report their self-assessed health as poor or less-than-good than higher income groups, at the same level of a more ‘objective’ health indictor such as the McMaster Health Utility Index. The degree of inequality in ‘subjective’ health is slightly higher than in ‘objective’ health, but not significantly different.

The degree of inequality in self-assessed health in Canada was found to be significantly higher than that reported by van Doorslaer et al. (1997, income related inequalities in health: some international comparisons, Journal of Health Economics 16, 93–112) for seven European countries, but not significantly different from the health inequality measured for the UK or the US. It also appears as if Canada’s health inequality is higher than what would be expected on the basis of its income inequality.

Introduction

An association between socioeconomic status (SES) and health has been observed for hundreds of years (Kaplan and Keil, 1993, Lynch et al., 1996). Regardless of the SES measure employed or the health outcome measured, the link between SES and health status is evident in our earliest records and exists in every country in which the relationship has been examined. Considerable work in this field, especially in Britain (Townsend and Davidson, 1982) and the rest of Europe (Fox, 1989, Kaplan and Keil, 1993, Lynch et al., 1996, Mackenbach et al., 1997, Kunst and Mackenbach, 1994), has led to the conclusion that lower socioeconomic status is associated with poorer health and that this association shows a graded rather than a threshold effect (Wilkinson, 1997).

There are several approaches to the measurement of socioeconomic status, including level of attained education, occupation, or household income (Mackenbach and Kunst, 1997). We have used income as the stratifying variable because it allows for comparison with a recent European study that employed the same methods. The use of household income as a measure of socioeconomic status is also consistent with Canadian health policy concerns — the reduction of income-related inequalities in health. “The first challenge we face is to find ways of reducing inequities in the health of low- versus high-income groups in Canada.” (Epp, 1986) Despite considerable work in the area of SES and morbidity, internationally, relatively little work has been reported in Canada (Cairney and Arnold, 1996, Mustard et al., 1997, Wilkins et al., 1989). Cairney and Arnold (1996) looked at socioeconomic determinants of self-assessed health and morbidity in elderly non-institutionalized Canadians. They demonstrated a strong inverse relationship between income and self-assessed health. Wilkins et al. (1989) also demonstrated an association between income and mortality by looking at deaths occurring to residents of urban centres in Canada during the period 1971–1986. Using a representative sample of the Manitoba population, Mustard et al. (1997) were able to show that mortality was inversely related to both education and income. However, a comprehensive investigation of the association between socioeconomic status and health status for Canada, across all regions and age groups, has not been accomplished to date.

The aim of this study is to provide evidence on income-related health inequalities in Canada using interview survey data from the 1994 National Population Health Survey. Our measure of health inequality, the ill-health concentration index, will be applied to two health status indicators. The first indicator is self-assessed health (SAH), a widely used indicator that is based on respondents’ own assessment of their health. The second indicator is the Health Utility Index (HUI) score, which is based on functional status assessed in eight domains. The HUI-based inequality measure will be used to check how it corresponds to the SAH-based measure and to test whether using the SAH or the HUI leads to different results. We will also compare the level of income-related health inequality in Canada to that reported for other OECD countries.

Section snippets

Methods

We have measured inequality in health using the ill-health concentration index proposed by Wagstaff et al. (1991). This measure meets three basic requirements for an index of inequality in health: (i) it reflects the socioeconomic dimension to inequalities in health; (ii) it reflects the experience of the entire population; and (iii) it is sensitive to changes in the distribution of the population across socioeconomic groups. The ill-health concentration index is defined as twice the area

Data and variable definitions

The survey we used in this analysis is the 1994 Canadian National Population Health Survey (NPHS), the first wave of a panel of individuals who will be followed longitudinally, at two-year intervals. The NPHS was designed to provide data for analytic studies that will assist in understanding the determinants of health, including the economic, social, demographic, occupational and environmental correlates of health. The target population of the NPHS includes household residents in all provinces,

Income-related health inequality

Table 1 provides the latent self-assessed ill-health scores deriving from the application of the standard lognormal assumption to the observed discrete distribution survey responses across the five categories of SAH. It can be seen that, as a result of the skewed distribution, the difference between the scores for, say, very good and good, is much smaller than the difference between the scores for fair and poor. The pattern is broadly consistent with the pattern observed for (1-HUI), which

Conclusion and discussion

The primary aim of this study was to test for the existence of nationwide income-related inequalities in health in Canada. Using data from the 1994 National Population Health Survey and applying the methods developed by Wagstaff and van Doorslaer (1994), we found evidence of relatively high income-related inequality in health in Canada. Our results clearly indicate that in Canada, as in most other countries, health inequalities favour the better-off.4

Acknowledgements

The authors would like to thank Cam Mustard, Mike Wolfson and Adam Wagstaff for helpful comments on a previous draft of this paper. They are also grateful to Statistics Canada for providing access to the data from the 1994 National Population Health Survey, and to Frank Puffer for supplying some additional computations from the 1987 US National Medical Expenditure Survey. EvD also wishes to acknowledge the generous hospitality of the Centre for Health Services and Policy Research at the

References (31)

  • J. Epp

    Achieving health for all: a framework for health promotion

    Canadian Journal of Public Health

    (1986)
  • S. Ettner

    New evidence on the relationship between income and health

    Journal of Health Economics

    (1996)
  • R. Evans et al.

    Why Are Some People Healthy and Others Not? The Determinants of Health of Populations

    (1994)
  • D. Feeny et al.

    Multi-attribute health status classification systems

    Pharmacoeconomics

    (1995)
  • J. Fox

    Health Inequalities in European Countries

    (1989)
  • Cited by (150)

    View all citing articles on Scopus
    View full text