Education level as a predictor of the onset of health problems among China's middle-aged population: Cox regression analysis

Abstract

Background: Despite the importance of midlife with reference to one's health, educational inequalities in midlife health have attracted little attention in China. Using Cox proportional hazards regression analysis, this study examined the association between educational attainment and the onset of midlife health problems and investigated the potential mediating effects of socioeconomic position (SEP) other than educational attainment, depression, and health behavior.

Methods: Data were extracted from the China Health and Retirement Longitudinal Survey (CHALRS) from 2011 (baseline) to 2018 (latest data). Participants aged 45-59 years at baseline were studied (N = 8,050). Health outcomes included the onset of poor self-rated health (SRH), limitation in activities of daily living (ADL) and instrumental ADL (IADL), multimorbidity, hypertension, dyslipidemia, heart diseases, and stroke over the 7-year follow-up period. Cox proportional hazard models were used to examine the associations of the outcomes with educational attainment, while controlling for potential mediators (other SEP, depression, and health behaviors).

Results: Lower educational level was associated with increased incidences of poor SRH and ADL/IADL limitations, but with decreased incidences of dyslipidemia and heart disease. After adjusting for baseline covariates, the RII was 2.17 (95% confidence interval [CI]: 1.74, 2.70) for poor SRH, 2.15 (95% CI: 1.42, 3.26) for ADL limitation, 3.84 (95% CI: 2.98, 4.94) for IADL limitation, 0.52 (95% CI: 0.40, 0.68) for dyslipidemia, and 0.55 (95% CI: 0.40, 0.74) for heart disease. Significant proportions (2.1 to 27.0%) of the RII were explained by the mediators. No sex or urban-rural differences were found in this study.

Conclusion: Our findings suggest that educational attainment is an important predictor of the incidences of key midlife health problems, with significant mediating effects exerted by other indicators of SEP, depression, and health behavior.

Keywords: Cox proportional hazard model; educational level; midlife health; non-communicable disease; relative index of inequality.