摘要
目的:构建学龄前恶性肿瘤患儿社会适应不良风险预测模型并进行验证,为提升其社会适应水平及远期生存质量提供依据。方法:选取7家三级甲等医院2021年7月—12月475例在院的学龄前恶性肿瘤患儿及主要照护者,以7∶3比例随机分为训练集和测试集。训练集通过社会适应评定量表筛查出社会适应不良患儿,运用Logistic回归和XGBoost方法确定预测因素并构建模型;测试集以受试者工作特征曲线(ROC)下面积、准确率、F1-分数、Kappa系数等指标验证模型的一致性和预测精度。结果:学龄前恶性肿瘤患儿社会适应不良的发生率为44.0%;XGBoost预测模型输出的社会适应不良风险因素重要性排序为一般家庭功能、年龄、平均网络使用时长、居住地、幼儿依恋关系、疾病阶段、家庭应对方式、是否复发。ROC曲线下面积为0.839,准确率为83.9%,Kappa系数为0.826,F1-分数为0.857。结论:构建的学龄前恶性肿瘤患儿社会适应不良风险预测模型具有良好的预测精度和一致性,可甄别个体社会适应不良风险。
Objective:To construct and validate the risk prediction model of social maladjustment in preschool children with malignancies.Methods:From July to December 2021,475 preschool children with malignancies and their primary caregivers were recruited.They were randomly assigned to training set(n=332)and test set(n=143)in a ratio of 7:3.The training set was screened for socially maladjusted children by the Social Adjustment Rating Scale,and Logistic regression and XGBoost methods were used to identify predictors and construct the model.The Receiver Operating Characteristic(ROC)curve,accuracy,F1-Score,and Kappa were used to evaluate the consistency and prediction accuracy of the model.Results:The incidence of the risk of social maladjustment in preschool children with malignancies was 44.0%.The XGBoost prediction model outputs the importance of maladjustment risk factors as general family functioning,age,the average length of internet use,region,infant attachment,stage of illness,family coping style,and whether it relapsed.The area under ROC curve of the model was 0.839,the accuracy was 83.9%,the Kappa was 0.826,and the F1-Score was 0.857.Conclusion:The prediction model based on XGBoost has good prediction accuracy and consistency,which can be used to screen individual adaptation maladjustment risk.
作者
刘艺璇
莫霖
沈雨晴
刘洋
余璐
LIU Yixuan;MO Lin;SHEN Yuqing;LIU Yang;YU Lu(Department of Nursing,Children's Hospital of Chongqing Medical University,Chongqing,400014,China)
出处
《中国护理管理》
CSCD
2023年第7期1014-1020,共7页
Chinese Nursing Management
基金
2022年重庆市自然科学基金面上项目(CSTB2022NSCQ-MSX0082)。
关键词
学龄前
儿童
恶性肿瘤
社会适应
预测模型
preschool
children
malignancies
social adjustment
prediction model
