基于一般临床资料与视觉质量评估结果构建初期皮质性白内障发病的Cox比例风险回归预测模型  

Construction of a Cox Proportional Risk Regression Prediction Model for the Development of Initial Cortical Cataract Based on General Clinical Data and Visual Quality Assessment Results

作　　者：王彦[1] 张立友 WANG Yan;ZHANG Li-you(Department of Cataract,Cangzhou Central Hospital/Cangzhou Eye Hospital,Cangzhou,Hebei 061000,China)

机构地区：[1]沧州市中心医院沧州眼科医院白内障科,河北沧州061000

出　　处：《临床误诊误治》2023年第1期124-128,共5页Clinical Misdiagnosis & Mistherapy

基　　金：河北省医学科学研究课题计划(20200343)。

摘　　要：目的分析初期皮质性白内障发病特征,基于一般临床资料与视觉质量评估结果构建Cox比例风险回归预测模型。方法选择2019年6月—2021年12月收治的初期皮质性白内障500例作为研究组,选取同期健康体检者500例作为对照组。分析初期皮质性白内障发病特征,对比2组一般临床资料及视觉质量评估结果,借助Cox回归分析筛选变量并构建Cox比例风险回归预测模型,根据受试者工作特征(ROC)曲线及曲线下面积(AUC)评估模型预测能力。结果500例初期皮质性白内障患者中,左眼占15.80%,右眼占17.00%,双眼占67.20%。女性白内障发生率高于男性(P<0.01),且随年龄增加,初期皮质性白内障占比逐渐增加。经Cox回归分析,初期皮质性白内障发病预测模型潜在预测因子为饮酒史、吸烟史、糖尿病史、冠心病史、对比敏感度、眩光敏感度、散射指数、体育锻炼频率。初期皮质性白内障发病预测模型的最终预测因子为饮酒史、吸烟史、糖尿病史、对比敏感度、眩光敏感度、散射指数、体育锻炼频率。Cox比例风险回归预测模型预测初期皮质性白内障发病的AUC为0.899,95%CI为0.897,0.917,敏感度为0.776,特异度为0.858,显示了良好的预测效果。结论基于一般临床资料与视觉质量评估结果的Cox比例风险回归预测模型能有效预测初期皮质性白内障发病风险。Objective To analyze the characteristics of initial cortical cataract onset,and to construct a Cox proportional risk regression prediction model based on general clinical data and visual quality assessment results.Methods A total of 500 patients with early cortical cataract admitted from June 2019 to December 2021 were selected as the research group,and 500 healthy physical examinees were selected as the control group during the same period.The characteristics of initial cortical cataract onset were statistically analyzed,the general clinical data and visual quality assessment results of the two groups were compared.The variables were screened and Cox proportional risk regression prediction model was constructed with the help of Cox regression analysis.The predictive ability of the model was assessed according to the receiver operating characteristic(ROC)curve and the area under the ROC curve(AUC).Results In 500 patients with early cortical cataract,the left eye accounted for 15.80%,the right eye accounted for 17.00%,and both eyes accounted for 67.20%.The incidence of cataracts in females was higher than that in males(P<0.01),and the proportion of early cortical cataracts increased with age.According to Cox regression analysis,the potential predictors of early cortical cataract were drinking history,smoking history,history of diabetes,history of coronary heart disease,contrast sensitivity,glare sensitivity,scattering index and physical exercise frequency.The final predictors of the model were drinking history,smoking history,history of diabetes,contrast sensitivity,glare sensitivity,scattering index,and physical exercise frequency.Cox proportional hazard regression model predicted the onset of early cortical cataract with an AUC of 0.899,a 95%CI of 0.897,0.917,a sensitivity of 0.776,and a specificity of 0.858,showing a good predictive effect.Conclusion A Cox proportional risk regression prediction model based on general clinical data and visual quality assessment results can effectively predict the risk of develo

关 键 词：皮质性白内障 对比敏感度 眩光敏感度 散射指数 预测模型 

分 类 号：R776.1[医药卫生—眼科]