The moderating effect of sociodemographic factors on the predictive power of self-rated health for mortality in Canada

James Falconer, Amélie Quesnel-Vallée


For decades the measurement of self-rated health has been a mainstay of health surveys. It has been consistently demonstrated as a reliable predictor for mortality. Yet, because a growing number of studies have found its predictive power to vary depending on social characteristics, debate still persists about the extent to which self-rated health can be validly used in the study of social inequalities in health. Using the Canadian context as an intermediate case between the many studies originating primarily from Europe and the United States, this study seeks to test the moderating effect of four sociodemographic characteristics on the power of self-rated health to predict mortality: age, sex, education, and income. Self-rated health, mortality, and a large set of control variables are measured using nationally-representative longitudinal panel data from the National Population Health Survey (NPHS), collected biennially from 1994-2010. Predictive power trajectories are modelled using time-series generalized estimating equation logistic regression. Findings show that self-rated health is a predictor for mortality up to 14 years prior to death in Canada. Despite their different overall levels of self-rated health, men and women show similar predictive power for mortality. Respondents over 80 showed reduced predictive power relative to respondents aged 30-64 or 65-79. However, testing the interaction between age and sex revealed that the age gradient in predictive power is only observable among women. Income and education only weakly moderated the predictive power of self-rated health in Canada, and the gradient was only observable between the highest and lowest education and income groups. This finding is consistent with the hypothesis that universal access to health care equalizes the availability of health information across the socioeconomic classes, and therefore diminishes the role of socioeconomic status for moderating the predictive power of self-rated health in Canada.


  • There are currently no refbacks.

Canadian Studies in Population | E-ISSN 1927-629X

Copyright © Canadian Population Society