REPORT SUMMARY

MANITOBA CENTRE FOR HEALTH POLICY

Socioeconomic Status and Health Care Services

How much is our health affected by our income and education? A wide-ranging study by the Manitoba Centre for Health Policy demonstrates that, throughout our lives but most strongly in midlife, the health of Manitobans varies markedly in relation to our socio-economic status.

It's not only that poverty and poor health are closely linked, as health care professionals have noted anecdotally for many years. This research demonstrates a persistent gradient in the occurrence of a range of illnesses at every level of socio-economic status; everyone is affected. People with the highest levels of education and income generally have better states of health than those with median education and income, who in turn are healthier than the poor and less educated. If this phenomenon were widely recognized, the report's authors stress, we might be able to develop the social commitment to address these inequalities and improve the quality of life for thousands of Manitobans.

 
A Unique Database

The first of three related research projects, this study used a unique database created through the collaboration of Statistics Canada, the Government of Manitoba and the University of Manitoba. It links the hospital and physician claims records for 48,000 Manitobans (a representative 5% sample of the province's population) with information they provided on their income and education levels from the 1986 Census. This complex dataset allows researchers to investigate the relationships between people's position in society and their health and use of health care services.

For each analysis, the sample population was divided into four groups or quartiles (Q) according to their level of education and household income. Q1 is the 25% of the sample with the lowest level of education or income; Q4 denotes the highest 25%.

The confidentiality of individuals has been protected throughout the creation and use of this database. Before researchers can handle the data, all names and addresses are excluded. Individuals are identified only by a scrambled version of their Manitoba Personal Health Information Number which cannot be matched to numbers actually in use in the health care system.

 
Death Rates and Socio-Economic Status

The risk of death was clearly associated with both income and education ranking. For every 100 people in the highest socio-economic group who died during a two-year period, approximately 140 deaths occurred in the poorest quartile and close to 160 in the second poorest group. This held true whether socio-economic status was measured by income or education level. The second highest income and education group had about 120 deaths for every 100 in the top quartile.

The graph in Figure 1 shows the gradient effect of these differences. Individuals with both top education levels and incomes (Q4) have the lowest mortality rate; for others, death rates tend to rise as income and education levels fall. When age groups were examined, lower income meant a higher risk of death especially for those aged 30-49 and 50-64. Poorer people are more likely to die younger.

 
Illness, Age, and Socioeconomic Status

This project examined how often people in the sample population received medical treatment over a one-year period for 15 categories of disorder, and matched these treatment rates with their socio-economic status. As with mortality, some patterns emerged. For many medical problems-some examples are diabetes, mental illness, and disorders of the digestive system-the risk of illness is greater for people with lower levels of education or income. For one category the reverse was true: higher socio-economic status meant people were more likely to be treated for disorders of the eye and ear.

Again, the pattern is most pronounced for people in the middle age groups: 15-29, 30-49, and 50-64 (Figure 2). The highest income group (Q4) is represented by the horizontal line at 1.0; the greater the risk of illness relative to this group, the higher the number. The gradient effect (the lower the income, the higher the risk) is easy to see, particularly for ages 15-74. This concentration of the health effects of socio-economic status in the midlife years raises questions about where policies to improve health should be directed: Are these differences the consequence of childhood experience? Or are they are more immediately determined by the circumstances of adult life? Fertility (rates of pregnancy) was also analyzed, and an interesting difference between age groups appeared. For younger women (ages 15-29), fertility rates are highest among those in the low-income, low-education group. For older women of childbearing age (30-49 years), the pattern is reversed: Women in the top income or education groups are the most likely to become pregnant at this time of life.

It is important to note that researchers on this project believe their estimates are conservative. No research method is perfect; compromises always have to be made in order to present complex information in a manageable way. The methods used in this study probably underestimate the differences which exist between the health of people in varying socio-economic groups. For example, the health of each population group is estimated in this study by how often people in that group see a doctor or go to hospital for a particular problem; it does not account for people who are ill but do not seek treatment, although we know that those people make up a relatively small percentage of Manitobans. In addition, the broad categories of illness used to allow the examination of a wide spectrum of disorders probably mask some important differences which would be apparent in a more specific study of, say, rheumatoid arthritis. This was intended to be a "big picture" study on the impact of socio-economic status on health, something which had not been done in Canada before. The full report contains numerous tables with much more detailed information than can be summarized here.

 
Socio-Esocio-economicconomic Status and Use of Health Services

In order to understand how use of our tax-funded health care dollars are distributed throughout the population, this project also tracked how people's use of hospital and physician services relates to their socio-economic status. In the earlier part of the study looking at overall treatment rates, hospital care and ambulatory care (non-hospital physician visits) were combined. When these two types of care are separated, their relationship to socio-economic status can be quite different.

Figure 3 shows that dollars spent on physician visits tends to increase with education, particularly within each income group. The income groups themselves do not show a consistent pattern. For use of short-stay hospital days (less than 60 days; Figure 4), the pattern is reversed and very clear. The number of hospital days used goes up substantially as both income and education levels decline. The better off you are, the less likely you are to be hospitalized.

Some health problems showed a more marked relationship to socio-economic status when hospital and physician care were separated. For example, the use of hospital days for the treatment of complications per 1,000 pregnant women in each income quartile was 26 days for Q1 (the lowest group), 22 days for Q2, 19 days for Q3, and 7 days for Q4 (the highest income group). This clear gradient pattern was identical for differences in education. The same was true for injury-related hospitalizations: Q1 (the lowest education group) used 85 days per 1,000 people, Q2 used 62 days, Q3 used 72, and Q4 only 50 days. On the other hand, ambulatory care (visits to physicians) for injuries showed no difference across income or education levels.

 
What Can Be Done?

Quantifying these relationships-long suspected in Canada and described in similar studies in other countries-is one step. Taking action to improve the health status of most Manitobans is another. This study does not attempt to explain the pathways by which socio-economic status affects health, but based on a reading of other work around the world, it does offer several policy options:

  • Consider directing an even greater share of health care services to lower socio-economic groups. It may be the case that health care services are not being used by persons in lower socio-economic quartiles in proportion to their need.
  • More aggressively target preventive medical and health services, especially in young adulthood. Too many current health promotion activities are passive and biased to middle and upper social class cultures. They frequently require access to material resources which people of lower means do not have.
  • Formulate explicit public policies addressing inequalities in health. Health policies which articulate the reduction in inequalities as a principal goal and which are combined with explicit target objectives offer one important mechanism for addressing this issue. This approach requires investments in research to determine what works and the orientation of incentives across the system to see that it happens.
  • The report also recognizes that there are substantial pressures on the finances of our publicly funded health care system which demand attention. The job of developing health targets (health promotion and disease prevention) must not be confused with the task of reforming health services (treatment of disease). These are two distinct public policy agendas, although they can share considerable common ground.

    Two future reports in this project will use the Statistics Canada/MCHP database to describe differences in health across occupational groups, and to compare how socio-economic data based on individuals and data on larger units such as neighbourhoods describe differences in health and health care use.

     

    Summary written by Amy Zierler, based on the report: Socioeconomic Gradients in Mortality and the Use of Health Care Services at Different Stages in the Life Course: by Cam Mustard, Shelley Derksen, Jean-Marie Berthelot, Michael Wolfson, Leslie L Roos and KC Carriere.

    Questions About This Report?
    Return to Report Summaries Directory
    Web page design by RJ Currie: rjcurrie@cc.umanitoba.ca