logistic回归联合ROC曲线评价IT-MAIS得分在预测0-36月龄婴幼儿听力损失程度中的诊断价值

Diagnostic value of IT-MAIS scores by logistic regression and ROC curve in predicting the degree of hearing loss in infants and toddlers aged 0-36 months

目的:比较正常听力和不同程度听力损失婴幼儿的早期语前听能发展轨迹,并探讨婴幼儿有意义听觉整合量表(IT-MAIS)得分在0-36月龄婴幼儿听力损失程度中的诊断价值。方法:收集2009年3月-2021年3月231名听力正常和611名不同程度听力损失患儿的临床资料。对不同程度听力损失患儿IT-MAIS得分随年龄的变化进行非线性回归拟合。分别构建IT-MAIS总得分、察觉得分和辨识得分与评估年龄联合的三种logistic回归模型,绘制ROC曲线,评价诊断效能。结果:轻、中、重、极重度患儿IT-MAIS得分随年龄的变化规律和正常听力婴幼儿的发育规律相似,均随年龄的增长而增长,且听力损失程度越严重,增长速率越缓慢,能达到的峰值越低。联合IT-MAIS总得分和评估年龄构建的logistic模型区分轻度及以上、中度及以上、重度及以上和极重度的AUC最佳,分别为0.827,0.889,0.948,0.946。对于极重度听力损失患儿的诊断效能最优,敏感度为89.6%,特异度为88.4%。听力损失程度越重,IT-MAIS的区分准确性越高,诊断效能越佳。结论:基于IT-MAIS量表构建的logistic模型联合ROC曲线法在鉴别0-36月龄婴幼儿听力损失程度诊断中有良好的诊断效能。当婴幼儿无法配合小儿行为测听、小儿测听结果不可靠、没有电生理条件时,IT-MAIS量表有望用于辅助预估婴幼儿听力损失程度,更加全面地了解患儿的听功能状态,为后续制定康复干预策略提供依据,具有一定的临床应用价值。

Objective: This study is inorder to compare the early prelingual auditory development trajectory of infants and toddlers with normal hearing and different degrees of hearing loss, and to explore the value of the Infant-Toddler Meaningful Auditory Integration Scale(IT-MAIS) scores in the diagnosis of hearing loss severity in infants and toddlers aged 0-36 months. Methods: Eight hundred and forty-two cases of infant-toddler from March 2009 to March 2021 were selected as participants, including 231 cases with normal hearing and 611 cases with hearing loss. The IT-MAIS scores of participants with different degrees of hearing loss were fitted with nonlinear regression with age. By respectively constructing three logistic regression models of IT-MAIS total scores, perception scores, discrimination scores and evaluation age, the ROC curve was drawn to evaluate the diagnostic efficacy. Results: The IT-MAIS scores of children with mild, moderate, severe, and profound hearing loss participants changes with age are similar to the development of infants with normal hearing, and they all increase with age. And the more severe the hearing loss, the slower the growth rate, and the lower the peak value that can be reached. The logistic model constructed by combining IT-MAIS total score and evaluation age is the best to distinguish mild and above, moderate and above, severe and above, and profound hearing loss, whose AUC are 0.827, 0.889, 0.948, 0.946. The diagnostic efficiency is the best for infant-toddlers with profound hearing loss, with a sensitivity of 89.6% and a specificity of 88.4%. The more severe the hearing loss, the higher the discrimination accuracy of IT-MAIS and the better the diagnostic efficiency. Conclusion: The logistic model based on the IT-MAIS scale and the ROC curve method have a good efficiency in the diagnosis of hearing loss severity in infants and toddlers aged 0-36 months. When the child cannot cooperate with behavioral audiometry, the results of behavioral audiometry are unreliable, and there is no electrophysiological condition, the IT-MAIS scale is expected to evaluate the degree of infants’ hearing loss as an auxiliary tool. It can understand the children’s auditory function state more and provides a basis for the subsequent formulation of rehabilitation intervention strategies with certain clinical application value.