中国における高校生の睡眠の質の低下のリスク予測モデルの開発と検証

Poor sleep quality is extremely harmful to the health and learning abilities of senior high school students. This issue has garnered significant societal attention. This study aimed to develop and validate a risk prediction model for identifying poor sleep quality among senior high school students in China, thereby enabling schools and parents to identify high-risk individuals and implement timely interventions. This study employed a cross-sectional design. Cluster sampling was employed to recruit participants from among senior high school students in China for the purpose of conducting a questionnaire survey from July to August 2021. The questionnaire included questions concerning demographic information, psychological status, lifestyle habits, and sleep status. We divided the data into training and validation sets at a 7:3 ratio. The logistic regression method was used to construct a prediction model, and the model was visualized using a nomogram. To evaluate the discrimination ability of the model, we utilized the area under the receiver operating characteristic curve. Calibration plots and the Hosmer-Lemeshow test were also used to evaluate the calibration of the model. Furthermore, decision curve analysis was used to assess its clinical practicality. This study included 4793 senior high school students, 24.2% of whom had poor sleep quality. Multivariate logistic regression analysis revealed that interpersonal sensitivity, anxiety, depression, high academic pressure, coffee consumption, alcohol consumption, smoking (including second-hand smoke), eating before bedtime, staying up late, a poor sleep environment, and prolonged use of hand-held electronic devices were risk factors for poor sleep quality among senior high school students. We used these factors to construct a predictive model and visualized it with a nomogram. The AUC values for the training and validation sets were 0.862 (95% CI = 0.847-0.876) and 0.853 (95% CI = 0.830-0.876), respectively. Additionally, the Hosmer-Lemeshow test values for the training and validation sets were P = 0.682 and P = 0.1859, respectively. The prediction model constructed in this research has good predictive performance. It will help schools identify high-risk groups with poor sleep quality and provides references for subsequent prevention and treatment.