Risk Factors for Mortality Among Mid-Aged and Older Chinese: A Prospective Cohort Study Using 9-Year Follow-Up of the China Health and Retirement Longitudinal Study

Background: To our knowledge, a comprehensive analysis of risk factors for mortality among mid-aged and older people in China based on a large nationally representative population study over an extended period has not yet been carried out. We investigated risk factors for mortality in the China Health and Retirement Study over 9 years. We aimed to investigate the associations between socioeconomic, demographic, psychosocial, behavioral, and health-related factors and all-cause mortality. Methods: Participants aged 45 and older were enrolled in the China Health and Retirement Longitudinal Study (CHARLS) from July 2011 to March 2012 following a multi-stage stratified sampling procedure to ensure national representativeness. The lower age limit of 45 was chosen so that CHARLS could evaluate the ageing process well before people become eligible for retirement Participants are followed up regularly, with the most recent assessments in 2020. The study is multi-disciplinary, including a wide-ranging set of 32 measures of health, health behavior, demographics, and socioeconomics. We used Cox proportional hazard modeling to study the independent and joint contribution of risk factors to 9-year all-cause mortality. Findings: Some 17,708 participants were included in the baseline sample; 586 were lost to follow-up, and we excluded 459 aged under 45 at the baseline, 24 whose dates of death were unknown, and 12 with missing marital status, education, or household registration status. Of the 16,627 (51.6% were women) included in this study, 2,352 (41.5% were women) died during a median follow-up of 4.9 years (IQR 2.67-6.83). In fully adjusted models, the following factors were independent risk factors for mortality: age (HR=13 .44 for 75+ relative to aged 45-49), male sex (HR=1.60), widowed or never married (HR=1.45), low education (HR=1.50 for illiterate relative to high school or above), below-median household per capita expenditure (HR=1.14), current (HR=1.32) and past smoking (HR=1.37), poor self-rated health (HR=1.48 relative to good health), any difficulty in basic activities of daily living (HR=1.26) or instrumental activities of daily living (HR=1.35), chronic lung disease (HR=1.22), hypertension (HR=1.34), diabetes (HR=1.37) and chronic kidney disease (HR=1.38), underweight (HR=1.45), all at the significance level of p<0.01. Doctor-diagnosed cancer (HR=1.61), asthma (HR=1.19), and general cognitive impairment (HR=1.16) were significant at p<0.05. The sex-stratified analysis indicated that doctor-diagnosed cancer was not significant for women and strengthened for men (HR=2.16, p<0.05), and that past smoking was a leading risk factor for men (HR=1.41, p<0.01) but not for women. When we excluded individuals with serious illness at baseline, the association with past smoking increased in magnitude (HR=1.55, p<0.01). Interpretation: Mortality in China is multi-faceted, and risk factors include demographic, socioeconomic, behavioral, psychosocial, and health factors. An integrated analysis of risk factors is more useful for policy-making than analyses of separate factors that can inflate the magnitude of associations. To further reduce mortality, China should increase access to quality health care services, improve the prevention, diagnosis and management of chronic diseases, continue to invest in education, occupational health and healthier home environments, and devote more efforts to tobacco control and ensure adequate nutrition among older people. Funding Information: National Institute on Aging, the Natural Science Foundation of China, the China Medical Board, the World Bank, and Peking University. Declaration of Interests: We declare no competing interests. Ethics Approval Statement: The study protocol was approved by the ethical review committee (IRB) of Peking University. Written informed consent was obtained from all study participants. Ethics approval for the data collection in CHARLS was obtained from the Biomedical Ethics Review Committee of Peking University (IRB00001052–11015 for biomarkers and IRB00001052–11014 for the survey).