Associations between socioeconomic status and prevalence, incidence and complications of type 2 diabetes in China

Abstract

INTRODUCTION: The prevalence of diabetes and other non-communicable diseases (NCDs) in China has increased markedly in the past few decades due to China’s recent economic development and urbanisation, which are contributing to socioeconomic and epidemiological transition. Socioeconomic status (SES) is an important determinant of health. Evidence from developed countries shows that, during the epidemiological transition NCDs arise initially in population sub-groups of high SES and then become more common in lower SES groups. It is well documented that SES is inversely associated with both diabetes prevalence and incidence in developed countries. However, the opposite has been found in some developing countries. This thesis aims to describe the associations between SES and prevalence, incidence and complications of type 2 diabetes in several contemporary Chinese populations.

METHODS: First, I conducted a systematic review of studies reporting the association between SES (measured by education, income or occupation) and prevalent type 2 diabetes in mainland China, Taiwan and Hong Kong. Second, I used baseline cross-sectional data from the China Health and Retirement Longitudinal Study (CHARLS) in 2011-2012 and China Kadoorie Biobank (CKB) in 2004-2008 to describe the association between SES (measured by educational level and household expenditure in CHARLS, and educational level and household income in CKB) and prevalent type 2 diabetes in mainland China. Then, I used prospective data from CKB (from 2004-2008 to 2013) to describe the association between SES and incident type 2 diabetes in mainland China. Last, I used prospective data from the Joint Asia Diabetes Evaluation (JADE) Program (from 2007 to 2017) to describe the association between SES (measured by educational level) and incident cardiovascular disease (CVD) and all-cause mortality in Hong Kong Chinese with type 2 diabetes. I used logistic regression models to describe the association between SES and prevalent type 2 diabetes, and Cox regression models to describe the association between SES and incidence or complications of type 2 diabetes.

RESULTS: Thirty-three English language studies and 22 Chinese language studies were included in the systematic review. The associations between SES, measured by education, income or occupation, and prevalent type 2 diabetes were inconsistent between studies.

The CHARLS study population included 4791 men and 5309 women. The odds ratios (OR) for prevalent diabetes were 1.53 (95% CI: 1.10, 2.15) in men and 0.91 (95% CI: 0.64, 1.28) in women for the highest compared to lowest educational level, and were 1.20 (95% CI: 0.95, 1.52) in men and 1.22 (95% CI: 0.98, 1.52) in women for the highest compared to lowest household expenditure group, after adjustment for age, residence, geographic areas and educational level or household expenditure as appropriate.

The cross-sectional analyses for CKB included 209352 men and 300867 women of whom 11616 men and 18450 women had diabetes. The OR for prevalent diabetes were 1.21 (95% CI: 1.09, 1.35) in men and 0.69 (95% CI: 0.63, 0.76) in women for the highest compared to lowest educational level, and were 1.45 (95% CI: 1.34, 1.56) in men and 1.26 (95% CI: 1.19, 1.34) in women for the highest compared to lowest household income group, after adjustment for age, study areas, family history of diabetes, educational level or household income as appropriate, and household size (for household income).

The prospective analyses for CKB included 197736 men and 282417 women without diabetes at baseline of whom 3607 men and 5937 women were identified as developing diabetes during follow-up. The hazard ratios (HR) for incident diabetes were 1.27 (95% CI: 1.07, 1.51) in men and 0.80 (95% CI: 0.67, 0.95) in women for the highest compared to lowest educational level, and were 1.36 (95% CI: 1.19, 1.55) in men and 1.06 (95% CI: 0.95, 1.17) in women for the highest compared to lowest household income group, after stratification by age (5-year age group) and study areas, and adjustment for age (continuous), family history of diabetes, educational level or household income as appropriate, and household size (for household income). The analyses of JADE data included 8934 men and 7479 women with type 2 diabetes. The HR for CVD and all-cause mortality were 0.74 (95% CI: 0.61, 0.93) and 0.65 (95% CI: 0.51, 0.84) in men, and were 0.69 (95% CI: 0.47, 1.02) and 0.42 (95% CI: 0.23, 0.76) in women for the highest compared to lowest educational level, after adjustment for age, diabetes duration and family history of diabetes.

CONCLUSIONS: Educational level was positively associated with prevalent and incident diabetes in Chinese men, while these associations were inverse in Chinese women in both CHARLS and CKB. Household income was positively associated with prevalent diabetes in both Chinese men and women in CHARLS. However, household income was positively associated with incident diabetes in men, but there was no significant association between household income and incident diabetes in women in CKB. In contrast, educational level was inversely associated with risk of CVD and all-cause mortality in both Hong Kong Chinese men and women with type 2 diabetes.

Information on socioeconomic inequalities in diabetes prevalence, incidence and complications in China can be used to plan health services, to develop approaches to primary and secondary prevention of diabetes and to monitor the epidemiological transition and the effects of policies designed to reduce socioeconomic and health inequalities.