县域经济发展与20~75岁人群糖尿病流行的 关联：基于多源数据的实证研究

Association between County-Level Socioeconomic Development and Diabetes Prevalence among Adults Aged 20~75: A Multi-Source Data Empirical Study

摘要: 目的：研究我国县域经济水平同20~75岁人群糖尿病发病风险之间的联系，目的在于给区域化糖尿病风险评价和慢性病防控策略的制定提供科学依据和数据支撑。方法：本文以2018年中国慢性病和危险因素监测数据、人口普查数据、卫生统计年鉴数据为基础，对284个县(区)中123个糖尿病监测点进行选择，共调查20~75岁居民58,705人。采用主成分分析法对县域经济发展中的15个主要指标进行维度简化和因子提取。采用两层随机效应逻辑回归模型，用个体作为第一层，用县级行政区划作为第二层来研究县域经济社会发展水平对糖尿病流行趋势的影响机制。所有的统计分析都是用SPSS 26.0软件进行的，采用MLwiN 2.30平台对多层次建模过程进行精确处理。结果：采用主成分分析法提取出五个对糖尿病流行风险有明显影响的公共因子，累积方差贡献率为80.46%，包含社会经济发展水平、城市基础设施建设情况、人口老龄化趋势、工业化发展程度和环境污染状况。从多层回归模型可以看出，在社会经济发展水平方面，高收入或者中等收入城市居民的糖尿病患病风险明显高于低收入城市居民(上分位数OR = 1.534，95%CI：1.229~1.776；中分位数OR = 1.232，95%CI：1.038~1.487)。同时区域环境污染物排放强度高组的居民糖尿病发病风险也明显增加，上分位数OR为1.311，95%CI为1.121~1.531。基础设施建设指标中高投入地区居民相关疾病发生率呈上升趋势。经多变量调整后，尽管部分因素如社会经济发展与污染暴露仍维持较高显著性(上分位数OR分别为1.247，95%CI：1.083~1.458；1.173，95%CI：1.017~1.355)，但其余因子未表现出统计学意义(P > 0.05)。结论：经济繁荣但是环境压力大的地区，居民患糖尿病的概率会明显上升。建议有关行政机构把这个地区当作慢性病管理的重点对象，健全健康监测体系并加以完善，改善干预措施的效果。

Abstract: Objective: By studying the relationship between the development status of county-level economy in China and the prevalence risk of diabetes among people aged 20~75, so as to provide a scientific basis for Regionally predict diabetes risk and chronic disease assessment strategy. Methods: According to data from the 2018 China Chronic disease and risk factor surveillance, national census and statistical yearbook, 123 monitoring sites of diabetes surveys covering 284 counties (districts) were selected as the study subjects. After filtering, 58,705 people aged 20~75 were included. Using principal component analysis to reduce the dimensions of 15 county-level economic development indicators and extracting common factors. Constructed a two-level logit random intercept model with individual as first-level unit, and counties/districts as second-level units to explore the influence of socioeconomic conditions in every county on the risk of diabetes prevalence. We did the statistic by means of SPSS 26.0 and MLwiN 2.30. Results: By principal components analysis, the top five factors (cumulative percentage: 80.46%) were extracted and consisted of socioeconomic development, urban infrastructure, population aging, industrial growth and pollution. The highest level of Socioeconomic development Tertile (OR = 1.534, 95%CI: 1.229~1.776), medium level Tertile (OR = 1.232, 95%CI: 1.038~1.487), pollution most Tertile (OR = 1.311, 95%CI: 1.121~1.531) are all associated positively with an increased risk of diabetes. Urban infrastructure also had a risk trend. After adjustment, only high socioeconomic development (OR = 1.247, 95%CI: 1.083~1.458) and high pollution (OR = 1.173, 95%CI: 1.017~1.355) were still risk factors. Other items were not significantly different (P > 0.05). Conclusion: The counties which grew economically more and had more pollution were likely to have diabetes. Thus, governments and other departments should prioritize the above areas in chronic disease prevention and control.