Duan Li, He Juan, Li Min, Dai Jiali, Zhou Yurong, Lai Feiya, Zhu Gang
Department of Psychiatry, The First Affiliated Hospital of China Medical University, Shenyang, China.
Central Laboratory, The First Affiliated Hospital of China Medical University, Shenyang, China.
Front Psychiatry. 2021 Jun 8;12:652356. doi: 10.3389/fpsyt.2021.652356. eCollection 2021.
Smartphone addiction has emerged as a major concern among children and adolescents over the past few decades and may be heightened by the outbreak of COVID-19, posing a threat to their physical and mental health. Then we aimed to develop a decision tree model as a screening tool for unrecognized smartphone addiction by conducting large sample investigation in mainland China. The data from cross-sectional investigation of smartphone addiction among children and adolescents in mainland China ( = 3,615) was used to build models of smartphone addiction by employing logistic regression, visualized nomogram, and decision tree analysis. Smartphone addiction was found in 849 (23.5%) of the 3,615 respondents. According to the results of logistic regression, nomogram, and decision tree analyses, Internet addiction, hours spend on smartphone during the epidemic, levels of clinical anxiety symptoms, fear of physical injury, and sex were used in predictive model of smartphone addiction among children and adolescents. The C-index of the final adjusted model of logistic regression was 0.804. The classification accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and AUC area of decision tree for detecting smartphone addiction were 87.3, 71.4, 92.1, 73.5, 91.4, and 0.884, respectively. It was found that the incidence of smartphone addiction among children and adolescents is significant during the epidemic. The decision tree model can be used to screen smartphone addiction among them. Findings of the five risk factors will help researchers and parents assess the risk of smartphone addiction quickly and easily.
在过去几十年中，智能手机成瘾已成为儿童和青少年中的一个主要问题，并且可能因新冠疫情的爆发而加剧，对他们的身心健康构成威胁。因此，我们旨在通过在中国大陆进行大样本调查，开发一种决策树模型作为未被识别的智能手机成瘾的筛查工具。利用中国大陆儿童和青少年智能手机成瘾横断面调查的数据（n = 3615），通过逻辑回归、可视化列线图和决策树分析来构建智能手机成瘾模型。在3615名受访者中，有849人（23.5%）被发现有智能手机成瘾问题。根据逻辑回归、列线图和决策树分析的结果，网络成瘾、疫情期间花在智能手机上的时间、临床焦虑症状水平、对身体受伤的恐惧以及性别被用于儿童和青少年智能手机成瘾的预测模型。逻辑回归最终调整模型的C指数为0.804。检测智能手机成瘾的决策树的分类准确率、敏感性、特异性、阳性预测值、阴性预测值和AUC面积分别为87.3、71.4、92.1、73.5、91.4和0.884。研究发现，疫情期间儿童和青少年智能手机成瘾的发生率很高。该决策树模型可用于筛查他们中的智能手机成瘾情况。这五个风险因素的研究结果将有助于研究人员和家长快速、轻松地评估智能手机成瘾的风险。
