PDF(Pubmed)
Abstract：
With an increasing prevalence of thyroid nodules globally, this study investigates the potential correlation between the use of Bluetooth headsets and the incidence of thyroid nodules, considering the cumulative effects of non-ionizing radiation (NIR) emitted by these devices. In this study, we analyzed 600 valid questionnaires from the WenJuanXing platform using Propensity Score Matching (PSM) and the XGBOOST model, supplemented by SHAP analysis, to assess the risk of thyroid nodules. PSM was utilized to balance baseline characteristic differences, thereby reducing bias. The XGBOOST model was then employed to predict risk factors, with model efficacy measured by the area under the Receiver Operating Characteristic (ROC) curve (AUC). SHAP analysis helped quantify and explain the impact of each feature on the prediction outcomes, identifying key risk factors. Initially, 600 valid questionnaires from the WenJuanXing platform underwent PSM processing, resulting in a matched dataset of 96 cases for modeling analysis. The AUC value of the XGBOOST model reached 0.95, demonstrating high accuracy in differentiating thyroid nodule risks. SHAP analysis revealed age and daily Bluetooth headset usage duration as the two most significant factors affecting thyroid nodule risk. Specifically, longer daily usage durations of Bluetooth headsets were strongly linked to an increased risk of developing thyroid nodules, as indicated by the SHAP analysis outcomes. Our study highlighted a significant impact relationship between prolonged Bluetooth headset use and increased thyroid nodule risk, emphasizing the importance of considering health impacts in the use of modern technology, especially for devices like Bluetooth headsets that are frequently used daily. Through precise model predictions and variable importance analysis, our research provides a scientific basis for the formulation of public health policies and personal health habit choices, suggesting that attention should be paid to the duration of Bluetooth headset use in daily life to reduce the potential risk of thyroid nodules. Future research should further investigate the biological mechanisms of this relationship and consider additional potential influencing factors to offer more comprehensive health guidance and preventive measures.
摘要：
随着全球甲状腺结节患病率的增加，这项研究调查了使用蓝牙耳机与甲状腺结节发生率之间的潜在相关性，考虑到这些设备发出的非电离辐射（NIR）的累积效应。在这项研究中,我们使用倾向得分匹配（PSM）和XGBOOST模型分析了来自WenJuanXing平台的600份有效问卷，辅以SHAP分析，评估甲状腺结节的风险。PSM用于平衡基线特征差异，从而减少偏差。然后采用XGBOOST模型来预测风险因素，模型功效由接收器工作特征(ROC)曲线(AUC)下面积测量。SHAP分析有助于量化和解释每个特征对预测结果的影响，确定关键风险因素。最初，来自文娟兴平台的600份有效问卷进行了PSM处理，产生96个案例的匹配数据集用于建模分析。XGBOOST模型的AUC值达到0.95，在区分甲状腺结节风险方面具有较高的准确性。SHAP分析显示，年龄和每日蓝牙耳机使用时间是影响甲状腺结节风险的两个最重要因素。具体来说,延长每天使用蓝牙耳机的持续时间与发生甲状腺结节的风险增加密切相关，如SHAP分析结果所示。我们的研究强调了长时间使用蓝牙耳机与甲状腺结节风险增加之间的显着影响关系，强调在使用现代技术时考虑健康影响的重要性，特别是对于经常使用的蓝牙耳机等设备。通过精确的模型预测和变量重要性分析,我们的研究为制定公共卫生政策和个人卫生习惯选择提供了科学依据，建议在日常生活中应注意蓝牙耳机的使用时间，以降低甲状腺结节的潜在风险。未来的研究应进一步探讨这种关系的生物学机制，并考虑其他潜在的影响因素，以提供更全面的健康指导和预防措施。
