基于贝叶斯网络模型河南省老年人高脂血症患病影响因素网络关系分析及患病风险预测

Network relationship analysis of the influencing factors and disease risk prediction for hyperlipidemia in the elderly population of Henan province based on a Bayesian network model

目的了解河南省老年人高脂血症患病影响因素的网络关系并预测其患病风险,为制定高脂血症的预防干预措施提供参考依据。方法于2022年7—12月采用多阶段随机抽样方法在河南省抽取123741名≥60岁老年人进行问卷调查、体格检查和实验室检查,并在多因素非条件logistic回归分析基础上采用最大最小爬山(MMHC)算法构建高脂血症贝叶斯网络模型,对当地老年人高脂血症患病影响因素的网络关系进行分析并预测其患病风险。结果河南省最终纳入分析的116091名老年人中,患高脂血症者33023例,高脂血症患病率为28.45%;患高总胆固醇(TC)血症、高甘油三酯(TG)血症、低高密度脂蛋白胆固醇(HDL-C)血症和高低密度脂蛋白胆固醇(LDL-C)血症者分别为9873、13738、14765和4338例,高TC血症、高TG血症、低HDL-C血症和高LDL-C血症的患病率分别为8.50%、11.83%、12.72%和3.74%;多因素非条件logistic回归分析结果显示,女性、腰高比(WHtR)≥0.5、体质指数(BMI)≥18.5、中心性肥胖、高血压和糖尿病是河南省老年人高脂血症患病的危险因素,年龄≥80岁和初中及以上文化程度是河南省老年人高脂血症患病的保护因素;贝叶斯网络模型分析结果显示,性别、BMI、高血压和糖尿病与高脂血症呈直接相关,年龄、文化程度、WHtR和中心性肥胖与高脂血症呈间接相关;风险推理结果显示,BMI≥28.0男性高血压老年人的高脂血症患病风险最高为41.6%,肥胖老年人高脂血症的患病风险为38.8%,高血压合并糖尿病老年人高脂血症的患病风险为37.0%。结论河南省老年人高脂血症患病率相对较低,性别、年龄、文化程度、BMI、WHtR、中心性肥胖、高血压和糖尿病为当地老年人高脂血症患病的主要影响因素,在高脂血症的防控过程中尤其应对BMI≥28.0的男性高血压老年人加以关注。

Objective To understand the network relationship of influencing factors for hyperlipidemia in the elderly population of Henan province and predict their disease risk,providing a reference basis for formulating prevention and intervention measures for hyperlipidemia.Methods From July to December 2022,a multi-stage random sampling method was used to select 123741 elderly individuals aged≥60 years old in Henan province for questionnaire surveys,physical examinations,and laboratory tests.Based on multivariate unconditional logistic regression analysis,a hyperlipidemia Bayesian network model was constructed using the Max-Min Hill-Climbing(MMHC)algorithm to analyze the network relationship of influencing factors for hyperlipidemia in the local elderly population and predict their disease risk.Results Among the 116091 elderly individuals in Henan province finally included in the analysis,33023 had hyperlipidemia,with a prevalence rate of 28.45%;9873,13738,14765,and 4338 individuals had high total cholesterol(TC),high triglyceride(TG),low high-density lipoprotein cholesterol(HDL-C),and high low-density lipoprotein cholesterol(LDL-C),respectively,with prevalence rates of 8.50%,11.83%,12.72%,and 3.74%,respectively.Multivariate unconditional logistic regression analysis showed that female gender,waist-to-height ratio(WHtR)≥0.5,body mass index(BMI)≥18.5,central obesity,hypertension,and diabetes were risk factors for hyperlipidemia in the elderly population of Henan province,while age≥80 years old and junior high school education or above were protective factors.Bayesian network model analysis showed that gender,BMI,hypertension,and diabetes were directly associated with hyperlipidemia,while age,education level,WHtR,and central obesity were indirectly associated.Risk inference results showed that the highest risk of hyperlipidemia was 41.6%for elderly male hypertensive individuals with BMI≥28.0,38.8%for obese elderly individuals,and 37.0%for the elderly individuals with hypertension and diabetes.Conclusion The prevalence of hyperlipidemia in the elderly population of Henan province was relatively low.Gender,age,education level,BMI,WHtR,central obesity,hypertension,and diabetes were the main influencing factors for hyperlipidemia in the local elderly population.In the prevention and control of hyperlipidemia,attention should be particularly paid to the elderly male hypertensive individuals with BMI≥28.0.