The mediating impact of lifestyle factors in the relationship between socioeconomic status and self-reported health in a Norwegian cohort of women

Abstract

Introduction: Information on self-reported health (SRH) with a simple question on “How do you rate your health?” provides sufficient information to be a reasonable proxy of health. Socioeconomic inequalities in health are often monitored via SRH. Individuals with better socioeconomic status (SES) more likely to rate their SRH as good. The part of the effect of SES (income level and education) operates through lifestyle factors along with the physical environment, social environment, social support, and psychological (cognitive, emotional, and social capabilities) development. The extent of the mediation on the relation between SES and SRH with the individual lifestyle factor’s impact will provide more insights into the relationship between SES and SRH through lifestyle factors. Material and methods: Data were extracted from the Norwegian Women and Cancer Study (NOWAC). After the inclusion/exclusion criteria, 53,941 participants were included in the analytical study sample. The descriptive statistics were presented for different education and income levels, using percentages, mean and standard deviation, and chi-square tests to test the difference. Logistic regression was used to establish the association between SES and SRH and reported with Odds Ratios (ORs) with 95% confidence interval (CI) using the statistical software SPSS, version 26. Mediation analysis was performed using the Medflex package in R applying a counterfactual approach, which was reported with ORs with 95% bootstrap CI along with the proportion of mediation. Results: The odds of reporting poor SRH were nearly half with a higher household income level compared to the lowest household income level. Furthermore, in the highest income level, the odds of poor SRH were lowered by 80% compared to the lowest household income level. The odds of reporting poor SRH with a high level of education compared to the lowest education level were about 35% less. With the highest education level, the odds of reporting poor SRH were 53% less compared to the lowest education group. About 20% of the association between income level and SRH was identified to go through an indirect path via lifestyle factors (smoking, body mass index (BMI), physical activity (PA)), whereas this was 20%-30% when investigating the model with education. Smoking and BMI contributed more to the indirect effect in the model with education, where PA contributed relatively less. Whereas in the model with income, PA had the greatest contribution, followed by BMI and smoking. The proportion of mediation by all the lifestyle factors (smoking, BMI, and PA) on the association between income level and SRH was statistically significant and quite similar across different income levels. There was an increasing gradient in the proportion of mediation across increasing education levels in the model with education and lifestyle factors (smoking, BMI, and PA). Conclusion: SES (education and income level) had a prominent and statistically significant effect on SRH. Better SRH was observed with both higher income levels and years of education. Household income level shows a more prominent effect than years of education on differentiating across groups of SES. Each lifestyle factors (smoking, BMI, and PA) individually and jointly mediated the association between SES and SRH. Any interventions focusing on improving the healthy lifestyle of the population should comprehend the impact of SES.