石化企业职业人群高血压危险因素分析及其患病风险预测

Risk factors analysis and risk prediction of hypertension in occupational population of petroleum enterprises

目的探讨石化企业职业人群高血压的影响因素,并建立高血压患病风险模型及个体预测工具。方法选取5799例石化企业职工为研究对象,以现场调查和职业健康检查资料为基础,分析高血压患病情况。采用多因素logistic回归分析筛选高血压相关影响因素;采用R软件构建高血压患病风险预测列线图。结果5799例研究对象平均年龄为(36.9±5.8)岁,高血压的患病率为13.17%。多因素logistic回归分析结果显示,年龄、性别、噪声暴露、血糖、甘油三酯、低密度脂蛋白、尿常规、心电图、肝胆B超和泌尿B超与石化企业职业人群高血压有统计学意义(OR=1.065、2.418、1.222、2.086、2.411、1.332、1.917、1.969、2.184、1.631,P<0.05)。利用独立影响因素构建的列线图具有良好的区分度(AUC:0.792;95%CI:0.776~0.809)。校准曲线提示具有较好的一致性。进行Boostrap内部交叉验证1000次,模型预测平均错误率为0.015,提示列线图稳定性较好。临床决策曲线显示,模型具有良好的临床实用性。结论基于石化企业职业人群高血压患病影响因素构建的列线图模型,具有良好的准确性和临床应用价值,可用于石化职业人群患高血压筛选,为个体化干预提供有价值的信息。

Objective To explore the influencing factors of hypertension among occupational population in petrochemical enterprises,and to establish the risk model of hypertension and individual prediction tools.Methods A total of 5799 cases of occupational population in petrochemical enterprises were selected as the research object,and the prevalence of hypertension was analyzed based on the field survey data and occupational health examination.Multivariate logistic regression analysis was used to screen the related influencing factors of hypertension.The risk prediction nomograph of hypertension was constructed with R software.Results The average age of 5799 subjects was(36.87±5.8)years,and the prevalence of hypertension was 13.17%.Multivariate Logistic regression showed that age,gender,noise exposure,blood glucose,triglyceride,low-density lipoprotein,the result of routine urine test,the result of electrocardiogram,the result of liver and gallbladder ultrasound and the result of urinary ultrasound were statistically significantly associated with hypertension in the occupational population of petrochemical enterprises(OR=1.065,2.418,1.222,2.086,2.411,1.332,1.917,1.969,2.184,1.631,P<0.05respectively.).The nomogram was constructed using the above 10 independent influencing factors,which has good discrimination(AUC:0.792;95%CI:0.776-0.809).The calibration curve indicated good consistency.After 1000 times of Bootstrap internal cross validation,the average error rate of model prediction was 0.015,indicating that the nomogram was stable.The clinical decision curve showed that the model had good clinical practicability.Conclusion The nomograph model constructed based on the influencing factors of hypertension in the occupational population of petrochemical enterprises has good accuracy and clinical application value,which can be used for screening hypertension in the occupational population of petrochemical enterprises,providing valuable information for individualized intervention.