Paper: Relative and absolute inequalities in cause-specific mortality among Canadian adults by socio-economic indicator

Author(s) and Affiliation(s):
Jacob Etches, Institute for Work & Health
Cameron Mustard, Institute for Work & Health / University of Toronto
Day/Time: Saturday at 14:00
Room: Ballroom, 2nd Floor

To describe the extent of social inequalities in mortality among adults in Canada by age, sex, cause of death, and socioeconomic status (SES) indicator, and to explore various relative and absolute methods of summarizing this inequality.


Ten-year mortality for a 15% sample of the adult population of Canada was collected by linking the 1991 Census long-form questionnaire to death certificates. Cox models were used to estimate the hazard of death by cause of death, sex, age and SES indicator for adults aged 30 to 69 in 1991. SES indicators include one measure of education, two measures of income and four measures of occupational rank. Causes of death were aggregated by chapter of the International Classification of Diseases, and also by two measures of medically amenable diseases, a measure of diseases due to alcohol and a measure of diseases due to tobacco. Summary measures include the Relative Index of Inequality (RII), the Population Attributable Risk %, the excess rate, excess death, and Potential Years of Life Lost (PYLL).


The RII is two to three for most causes of death and SES indicators. Relative inequality was slightly greater for men, and decreased with age for most causes of death. Relative inequality varied by cause of death but rarely by socioeconomic indicator. Excess rates are high for common causes of death, men, and at older ages, and are mostly consistent across socioeconomic indicators. Excess deaths are greater at ages 50-59 than 60-69 due to the greater population at risk. PYLL are greater for men, and show cause-of-death-specific age patterns.


There are large socio-economic inequalities in mortality among Canadian adults. Instances in which the extent of inequality varies importantly by socioeconomic indicator are rare. Men experience more inequality than women according to relative, and especially absolute measures. Magnitudes of inequality differ by cause of death and age, but the pattern depends importantly on the summary measure used, even among absolute measures. The appropriate measure depends on the purpose of the analysis, the audience, and on value-judgements regarding what constitutes inequality.