论文部分内容阅读
目的构建50岁以上健康管理队列的白内障发病风险预测模型。方法依托山东多中心健康管理纵向观察数据库,采用Cox比例风险回归构建白内障发病风险预测模型,通过ROC曲线下面积(AUC)评价模型的预测效果,并利用十折交叉验证来检验模型的稳定性。结果随访期间共新发白内障病例1 010例,发病密度为24.76‰。预测模型最终纳入年龄、性别、吸烟、高黏稠血症、鼓膜疾患、屈光不正、糖尿病、总胆固醇和收缩压9个变量。白内障发病风险预测模型的AUC为0.712(95%CI:0.693~0.732)。十折交叉验证的平均AUC为0.714。结论研究构建的白内障发病风险预测模型有较好的预测效果,为白内障高危人群的早期筛查提供了依据。
Objective To construct a predictive model for the risk of cataract in patients over 50 years old with health management cohort. Methods Based on the Shandong Multi-Center Health Management Longitudinal Observation Database, Cox proportional hazards regression was used to construct the risk prediction model of cataract. The area under the ROC curve (AUC) was used to evaluate the predictive value of the model. Ten-fold cross validation was used to test the stability of the model. Results During the follow-up period, a total of 1 010 new cataract cases were found, with a disease incidence of 24.76 ‰. The predictive model eventually included nine variables, age, gender, smoking, hyperviscosity, tympanic membrane disease, refractive errors, diabetes mellitus, total cholesterol and systolic blood pressure. The AUC of the cataract risk prediction model was 0.712 (95% CI: 0.693-0.732). The average AUC of ten-fold cross-validation was 0.714. Conclusions The prediction model of the risk of cataract may provide a good basis for the early screening of high-risk cataract patients.