基于健康管理队列的高血压风险预测模型

发布时间：2018-06-25 23:51

本文选题：高血压 + 健康管理队列　；参考：《山东大学学报(医学版)》2017年06期

【摘要】：目的基于健康管理队列,构建高血压风险预测模型。方法依托山东多中心健康管理纵向观察队列,排除基线高血压、心脑血管疾病、血肌酐177μmol/L、年龄20岁者,构建高血压研究队列(共22 177人,其中男12 044人,女10 133人),分性别采用Cox回归建立高血压预测模型,并评价模型预测效果。结果观察期间新发高血压4 571例,发病密度为62.84/1 000人年。最终男性模型中的变量包括年龄、体质量指数、收缩压、舒张压、空腹血糖和红细胞压积,ROC曲线下面积(AUC)为0.821(95%CI:0.812~0.830);女性模型中的变量包括年龄、体质量指数、收缩压、红细胞计数和高密度脂蛋白,AUC为0.818(95%CI:0.806~0.828)。十折交叉验证结果显示,男女AUC分别为0.819(95%CI:0.810~0.828)、0.814(95%CI:0.803~0.825)。结论该模型具有较好的预测能力,可用来识别高血压高危个体。
[Abstract]:Objective to construct a risk prediction model for hypertension based on health management cohort. Methods based on the longitudinal observation cohort of Shandong multi-center health management, we excluded baseline hypertension, cardio-cerebrovascular disease, serum creatinine 177 渭 mol / L, aged 20 years, and constructed a study cohort of hypertension (22 177 subjects, including 12 044 males). Female 10 133). Cox regression was used to establish the prediction model of hypertension, and the predictive effect of the model was evaluated. Results during the observation period, 4 571 cases of new hypertension were observed, the incidence density was 62.84 / 1 000 person-years. The variables in the final male model included age, body mass index, systolic blood pressure, diastolic blood pressure, fasting blood glucose and hematocrit under the ROC curve (AUC) of 0.821 (95 CI: 0.8120.830), and the female model included age, body mass index, systolic blood pressure. The red blood cell count and high density lipoprotein (HDL) AUC were 0.818 (95 CI: 0.806 / 0.828). The results of ten fold cross test showed that the AUC of men and women were 0.819 (95: 0.810 / 0.828) and 0.814 (95 CI: 0.803 / 0.825) respectively. Conclusion the model has good predictive ability and can be used to identify high-risk individuals with hypertension.
【作者单位】：山东大学公共卫生学院生物统计学系;山东大学齐鲁生物医学大数据研究中心;胜利油田中心医院保健科;胜利油田中心医院健康管理中心;山东大学附属千佛山医院健康管理中心;
【基金】：国家自然科学基金(81273082)
【分类号】：R544.1

【相似文献】