Explaining the social gradient in health in Canada: Cross-sectional and longitudinal analyses examining the role of stressors using the National Population Health Survey

Abstract

The pervasive relationship between socioeconomic status and health has been observed in virtually all Western countries, Canada notwithstanding. The relationship between higher socioeconomic status (SES) and better health has been demonstrated to be a stepwise gradient with better health at each successive level of SES, indicating that factors beyond absolute material poverty are likely to be causing this gradient. In order to attenuate social gradients in health, underlying mechanisms must be elucidated. The purpose of this thesis was to examine whether psychosocial stressors associated with lower SES explain the poorer health of poorer people. A secondary purpose of this thesis was to examine a health behaviour-mediated pathway between stressors and health. Cross-sectional analyses of the 1994-95 National Population Health Survey (NPHS), and longitudinal analyses of the 1994-95 and 1996-97 NPHS were conducted. Logistic regression analyses demonstrated that lower income was associated with a greater odds of concomitant fair/poor self-rated health, and with a greater odds of experiencing a decline in self-rated health over a two-year period. Most stressors were more prevalent among lower income groups, and with one exception, all stressors were associated with a greater odds of fair/poor self-rated health and of experiencing a decline in health status. Stressors mediated a modest proportion of the social gradient in health. In cross-sectional analyses, about a quarter of the relationship between income and health was explained for men, and from 6% to 15% for women. In longitudinal analyses, these stressors explained 16% and 10% of the relationship between the lowest and second lowest income quintiles and decline in self-rated health respectively. Furthermore, all stressors were associated with smoking behaviour, while fewer associations were observed between stressors and physical activity behaviour. Health behaviours mediated a modest proportion of the relationship between stressors and fair/poor health. The results of this thesis provide evidence for partial mediation of the social gradient in health by psychosocial factors, namely stressors. Furthermore, stress-related health behaviours may be a secondary pathway in this relationship. Future research is warranted to further refine explanatory models of the social gradient in health, in order to identify appropriate points for intervention.