2012-2021年江苏省＜15岁儿童急性弛缓性麻痹病例残留麻痹的影响因素和风险预测模型构建
Risk prediction and influencing factors of residual paralysis in acute flaccid paralysis cases under 15 years of age in Jiangsu province,2012-2021
李梅 1许燕 1陈强 1孙翔 1贾成梅 1冷红英 1吴昀 1郭宏雄 1汪志国1
作者信息
- 1. 江苏省疾病预防控制中心,江苏南京 210009
- 折叠
摘要
目的 分析急性弛缓性麻痹(Acute flaccid paralysis,AFP)病例发生残留麻痹的影响因素并构建风险预测模型.方法 通过江苏省AFP病例监测系统收集2012-2021年江苏省＜15岁儿童AFP病例,分析初始麻痹后60 d残留麻痹比例和影响因素,构建影响因素的风险预测模型并估计风险概率;采用受试者工作特征曲线(Receiver operating characteristic curve,ROC)、Bootstrap法和校准曲线评估模型预测效能.结果 2012-2021年江苏省＜15岁儿童AFP病例的残留麻痹比例为17.78％(367/2 064),多因素Logistic回归分析显示,年龄＜1岁、就诊次数≥2次、脊灰疫苗免疫史＜3剂次、发病至就诊时间≥7 d、初始麻痹前发热、麻痹期呼吸困难和大小便失禁病例的残留麻痹比例高.构建的风险预测模型显示各因素的残留麻痹风险概率在0.08-0.22之间;模型的ROC曲线下面积为0.745(95％CI:0.715-0.774),Bootstrap法计算的C-index值为0.740模型的校准曲线和Hosmer-Lemeshow检验显示预测风险概率与实际风险概率基本吻合.结论 本研究显示AFP病例发生残留麻痹主要与脊灰疫苗免疫史、年龄、就诊情况和部分发病症状等因素有关,构建的风险预测模型具有较好的预测效能.
Abstract
Objective To explore factors influencing residual paralysis in acute flaccid paralysis(AFP)cases and develop a risk prediction model for residual paralysis.Methods We obtained data on AFP cases under 15 years old reported to the Jiangsu AFP Case Surveillance System during 2012-2021 to determine the proportion with residual paralysis 60 days after initial paralysis and factors influencing residual paralysis.We constructed a prediction model to estimate the risk probability of residual paralysis and evaluated the model using receiver operating characteristic curve(ROC)analysis,a bootstrap method,and calibration curve analysis.Results The proportion of AFP cases＜15 years old with residual paralysis was 17.78％(367/2 064).Multivariate logistic regression analysis showed that the proportion was highest among infants and individuals with ≥2 doctor visits,who received＜3 doses of polio vaccine,who visited doctors ≥ 7 days after onset,with fever before onset,who had dyspnea during paralysis,and who had incontinence during paralysis.The risk prediction model showed that risks of residual paralysis for thesefactors rangedfrom 0.08 to 0.22,with an area under the ROC curve of 0.745(95％CI;0.715-0.774)and a C-index of 0.740.The calibration curve and Hosmer-Lemeshow tests found that the predicted risk probabilities of residual paralysis were comparable with actual values.Conclusions Residual paralysis in AFP cases was significantly associated with polio vaccination history,age,medical visits,and symptoms.A constructed risk prediction model had good prediction efficiency.
关键词
急性弛缓性麻痹/残留麻痹/影响因素/风险预测模型Key words
Acute flaccid paralysis/Residual paralysis/Influencing factor/Risk prediction model引用本文复制引用
基金项目
江苏省科技厅社会发展重点项目(BE2015714)
出版年
2024
NETL
NSTL
万方数据