基于体检队列的胃炎风险预测模型

A risk prediction model of gastritis based on examination cohort

目的构建基于体检者的胃炎风险预测模型。方法在2004年5月到2016年9月山东多中心健康管理纵向观测队列中选取初次体检未诊断为胃炎的体检者33 416人,通过多元逐步回归进行变量选择,建立Cox比例风险模型。最后通过十折交叉法进行内部验证。结果在观察期间有842例体检者出现新发胃炎情况,发病率为2.52%,发病密度为18.74‰,最后纳入模型的变量包括:年龄、幽门螺旋杆菌感染、白蛋白水平、嗜碱性粒细胞百分数以及淋巴细胞计数。预测模型的ROC曲线的AUC值为0.691(95%CI:0.670~0.712)。模型十折交叉验证后的AUC值为0.673(95%CI:0.652~0.694)。结论建立基于体检队列的胃炎风险模型的预测能力较好。

Objective To construct a risk prediction model based on the physical examination data.Methods A total of 33 416 non-gastritic subjects who received their first physical examination during May 2004 and Sep.2016 in multicenters were involved.The Cox proportional risk model was adopted to perform variable selection by stepwise method.Ten-fold cross validation was used to test the stability of the model.Results In the observation period,842 newgastritis cases were observed.The incidence rate was 2.52%and incidence density was 18.74‰.The variables included age,helicobacter pylori infection,albumin,basophils percentage and lymphocyte count.In the receiver operating characteristic(ROC)curve,the area under the curve(AUC)was 0.691(95%CI:0.671-0.710).The AUC after the ten-fold cross-validation was 0.673(95%CI:0.652-0.694).Conclusion The model we constructed can effectively predict the gastritis risk.