基于LASSO回归分析氟保护漆预防儿童龋齿效果的危险因素并构建预测模型OA
Based on LASSO regression to analyze risk factors for the effectiveness of fluoride varnish in preventing dental caries in children and to construct a prediction model
目的:基于 LASSO 回归分析氟保护漆预防儿童龋齿效果的危险因素,并建立预测模型.方法:回顾性分析 2020 年 3 月～2024 年 5 月在本院应用氟保护漆防龋的 278 例儿童资料,依据电子病历系统记录的干预 12 个月后龋齿发生情况分为发生组(n=134)和未发生组(n=144),收集并比较两组资料差异,采取 LASSO-Logistic 回归分析氟保护漆预防儿童龋齿效果的相关因素,利用 R 软件建立预测模型并评估其预测效能.结果:278 例儿童中发生龋齿 134 例(占 48.20%),未发生龋齿 144 例(占 51.80%).发生组与未发生组年龄、夜间进食频率、甜食食用频率、刷牙频率、单次刷牙时间、牙膏含氟、定期口腔检查及监护人口腔知识水平比较有统计学差异(P<0.05).LASSO 回归筛选出的变量纳入多因素 Logistic 回归分析,年龄(OR=2.108,P=0.017)、夜间进食频率(OR=3.283,P=0.001)、甜食食用频率(OR=2.207,P=0.002)、刷牙频率(OR=1.673,P=0.034)、牙膏含氟情况(OR=4.217,P<0.001)及监护人口腔知识水平(OR=2.995,P<0.001)均是氟保护漆预防儿童龋齿效果的危险因素.建立列线图模型,绘制受试者工作特征(receiver operating characteristic,ROC)曲线的曲线下面积(the area under the curve,AUC)为0.751(95%CI 0.696～0.806),校准曲线接近理想状态,且 Hosmer-Lemeshow 拟合优度检验 χ2=11.634,P=0.618.结论:氟保护漆预防儿童龋齿效果的危险因素包含年龄、夜间进食频率、甜食食用频率、刷牙频率、牙膏含氟及监护人口腔知识水平,基于上述 6 项因素建立的列线图模型对氟保护漆防龋效果具有良好预测价值.
Objective:To analyze the risk factors affecting the effectiveness of fluoride varnish in preventing dental caries in children based on LASSO regression and to establish a prediction model.Methods:A retrospective analysis was con-ducted on the data of 278 children who received fluoride varnish for caries prevention in our hospital from March 2020 to May 2024.The children were divided into two groups based on the occurrence of dental caries recorded in the electronic medical record system 12 months after intervention:the occurrence group(n=134)and the non-occurrence group(n=144).Data from both groups were collected and compared for differences.LASSO-Logistic regression was used to analyze factors related to the effectiveness of fluoride varnish in preventing dental caries in children.R software was employed to establish a predic-tion model and evaluate its predictive performance.Results:Among the 278 children,134(48.20%)had dental caries,and 144(51.80%)had no dental caries.There were significant differences in age,frequency of eating at night,frequency of eating sweets,frequency of brushing teeth,time of single brushing teeth,fluoride content in toothpaste,regular oral examination and oral knowledge level of guardians between the occurrence group and the non-occurrence group(P<0.05).The variables se-lected by LASSO regression were included in multivariate Logistic regression analysis.Age(OR=2.108,P=0.017),frequen-cy of eating at night(OR=3.283,P=0.001),frequency of eating sweets(OR=2.207,P=0.002),frequency of brushing teeth(OR=1.673,P=0.034),fluoride content in toothpaste(OR=4.217,P<0.001)and oral knowledge level of guardians(OR=2.995,P<0.001)were all risk factors for the effect of fluoride varnish on preventing dental caries in children.A no-mogram model was established,and the area under the curve(AUC)of the receiver operating characteristic(ROC)curve was 0.751(95%CI 0.696～0.806).The calibration curve was close to the ideal state,and the Hosmer-Lemeshow goodness of fit test was χ2=11.634,P=0.618.Conclusion:The risk factors for the effect of fluoride varnish on preventing dental caries in children include age,frequency of eating at night,frequency of eating sweets,frequency of brushing teeth,fluoride content in toothpaste and oral knowledge level of guardians.The nomogram model based on the above six factors has a good predictive value for the effect of fluoride varnish on preventing dental caries.
李晓林;余蕾;王北川;刘仲强
自贡市第四人民医院口腔科 四川 自贡 643000自贡市第四人民医院口腔科 四川 自贡 643000自贡市第四人民医院口腔科 四川 自贡 643000自贡市第四人民医院口腔科 四川 自贡 643000
医药卫生
氟保护漆龋齿危险因素预测模型
Fluoride varnishDental cariesRisk factorsPrediction model
《临床口腔医学杂志》 2026 (5)
279-283,5
评论