基于多重填补法合并广义估计方程的高血糖危险因素评估研究

马庆华; 毛建良; 徐文新; 孙宏鹏; 劳雅琴

doi:10.16766/j.cnki.issn.1674-4152.000574

基于多重填补法合并广义估计方程的高血糖危险因素评估研究

doi: 10.16766/j.cnki.issn.1674-4152.000574

1. 苏州市相城区第三人民医院预防保健科, 江苏苏州 215134;
2. 苏州大学医学部公共卫生学院儿少卫生与社会医学教研室, 江苏苏州 215123;
3. 苏州市相城区妇幼保健所, 江苏苏州 215131

基金项目:

国家自然科学基金项目(81402761)；江苏省自然科学基金项目(BK20140361)；苏州市相城区科技计划项目(XJ201540，XJ201706)

详细信息

通讯作者:
劳雅琴,E-mail:382303294@qq.com

中图分类号: R587.1;R151.2
计量
- 文章访问数: 243
- HTML全文浏览量: 42
- PDF下载量: 2
- 被引次数: 0
出版历程
- 收稿日期: 2018-01-18
- 网络出版日期: 2022-08-06

Research about the risk assessment of hyperglycemia based on multiple imputation and generalized estimating equation

Department of Preventive Medicine, the Third People's Hospital of Xiangcheng District, Suzhou, Jiangsu 215134, China

摘要

摘要: 目的应用多重填补法合并广义估计方程评估高血糖危险因素。方法利用多重填补法合并广义估计方程分别处理苏州市相城区第三人民医院老年人2011-2015年(2013年除外)健康管理档案中相关信息的数据缺失和重复测量问题,综合评估高血糖的危险因素。结果 8 325位老年人累计23 195条观测记录中,纳入的指标除性别、年龄为完整变量,其余变量缺失率范围为0.06%~18.44%,样本实际有效率为76.99%,数据资料呈任意缺失模式。多重填补后,产生10个完整的不同数据集,所有变量填补效率都在0.97以上。高血糖危险因素OR值(95% CI)评估结果:高血压1.272(1.201~1.361)、超重1.251(1.162~1.341)、肥胖1.649(1.481~1.852)、心率过快1.679(1.581~1.792)、高胆固醇血症1.178(1.060~1.311)、混合型高脂血症1.170(1.021~1.332)、低密度脂蛋白胆固醇1.112(1.040~1.171)、高尿酸血症1.172(1.089~1.271)、丙氨酸氨基转移酶偏高1.180(1.032~1.351)、天冬氨酸氨基转移酶偏高1.191(1.032~1.391)。缺失填补前后,混合型高脂血症和天冬氨酸氨基转移酶偏高的统计学意义发生改变。γ和r值表明缺失对体质指数的参数估计影响最大。结论多重填补法合并广义估计方程能有效解决高血糖危险因素评估时遇到的数据缺失和重复测量问题,公共卫生部门应加强监测可能引起高血糖患病的相关因素,针对高危老年人做好健康指导工作。
- 纵向数据 /
- 多重填补法 /
- 广义估计方程 /
- 高血糖 /
- 危险因素评估
Abstract: Objective To apply multiple imputation and generalized estimating equation to the risk assessment of hyperglycemia. Methods On the basic of the elderly health management archives data from The 3rd People's Hospital of Xiangcheng District from 2011 to 2015(except 2013), multiple imputation was used to solve the data missing problem and generalized estimating equation was used to solve the repeated measurements problem. Finally, comprehensive results of risk factors for hyperglycemia would be calculated. Results This study gathered 8 325 different elderly health management archives data and accumulated 23 195 observational data. In all contained indicators, there were only variables as gender and age without missing. The ranges of miss rate of the rest of variables were 0.06% to 18.44% and the actual effective rate of the sample was 76.99%. All the data sets had an arbitrary missing pattern, and 10 different complete data sets were produced after using multiple imputation. All variables' relative efficiency was above 0.97. Hyperglycemia multi-factor comprehensive inference showed that the OR values (95% CI):high blood pressure 1.272(1.201-1.361), overweight 1.251(1.162-1.341), obesity 1.649(1.481-1.852), a racing heart 1.679(1.581-1.792), hypercholesterolemia 1.178(1.060-1.311), combined-hyperlipidemia 1.170(1.021-1.332), low density lipoprotein cholesterol 1.112(1.040-1.171), high blood uric acid 1.172(1.089-1.271), high alanine transaminase 1.180(1.032-1.351), high aspartate transaminase 1.191(1.032-1.391). After the multiple imputation, the statistical significance of combined hyperlipidemia and high aspartate transaminase changed. The γ and r values showed that missing data had the greatest impact on the parameter estimation of body mass index. Conclusion The problems of data missing and repeated measurements can be solved effectively by multiple imputation and generalized estimating equation when assessing the risk factor for hyperglycemia, the departments of public health should strengthen the surveillance for related factors which may cause the hyperglycemia and enhance health education aiming at the elderly.
- Longitudinal data /
- Multiple imputation /
- Generalized estimating equation /
- Hyperglycemia /
- Risk assessment