【摘要】：目的建立20歲以上健康管理人群高脂血癥風險預測模型并對其預測效果進行評價。方法依托山東多中心健康管理縱向觀察隊列共納入30 056人,采用Cox比例風險回歸建立高脂血癥預測模型,利用ROC曲線下面積(AUC)進行模型評價,十折交叉驗證法檢驗模型的預測效果和判別能力。結(jié)果隨訪期間共新發(fā)高脂血癥5 063例,發(fā)病密度為47.78‰。預測模型納入的變量為年齡、性別、吸煙、飲酒、總膽固醇、甘油三酯、總膽紅素、高密度脂蛋白、糖尿病和高血壓10個變量。預測模型的ROC曲線下面積AUC為0.741(95%CI:0.731~0.752),經(jīng)十折交叉驗證平均AUC為0.741。結(jié)論構(gòu)建的高脂血癥風險預測模型在健康管理人群中具有較好預測能力。
[Abstract]:Objective to establish a risk prediction model for hyperlipidemia in health management population over 20 years old and evaluate its effect. Methods A total of 30 056 people were enrolled in the longitudinal observation cohort based on multi-center health management in Shandong Province. The prediction model of hyperlipidemia was established by Cox proportional risk regression, and the model was evaluated by area (AUC) under the ROC curve. The prediction effect and discriminant ability of the model were tested by ten fold cross validation method. Results during the follow-up period, there were 5 063 cases of new hyperlipidemia, the incidence density was 47.78 鈥，

本文編號：2188141