拒學青少年的睡眠模式與心理健康特質：來自腕動計、憂鬱量表與兒童行為量表之探討

Abstract

背景：拒學行為為一種常見於青少年的功能性缺課表現，伴隨顯著的焦慮、憂鬱與身體化症狀。儘管臨床觀察指出其與睡眠問題高度相關，過往研究大多依賴主觀報告，缺乏客觀睡眠資料的實證支持。 方法：本研究招募90位13至18歲青少年，依學校出席率與精神科診斷分為拒學組（SR）、焦慮／憂鬱組（AD）與健康對照組（HC）。使用腕動計（actigraphy）連續紀錄7日睡眠資料，取得總睡眠時間、睡眠後覺醒時間（WASO）、睡眠效率與睡眠規律指數（SRI）。心理症狀評估採用 CES-D 與 CBCL DSM導向量表。統計方法包含變異數分析、事後比較與邏輯斯迴歸。 結果：拒學組在SRI、WASO與睡眠效率上均顯著劣於其他兩組，且CES-D與CBCL得分亦明顯較高。邏輯斯迴歸指出，SRI較低以及身體化症狀較高為預測拒學行為的兩項重要指標。 結論：本研究提供首度使用客觀睡眠指標系統性驗證拒學行為與睡眠規律失調間關聯之證據。結果支持將睡眠規律性納入拒學風險評估與介入策略，尤其建議在傳統焦慮與憂鬱治療之外，進一步整合睡眠節律重建與行為節律支持之介入模式。

Background: School refusal in adolescents is a complex and heterogeneous phenomenon, often accompanied by anxiety and depressive symptoms. Although sleep disturbances are frequently observed in clinical settings, few studies have incorporated objective sleep assessments in this population. Methods: A total of 90 adolescents aged 13 to 18 were recruited and categorized into three groups: school refusal (SR), anxiety/depression (AD), and healthy controls (HC), based on clinical and school attendance records. Psychological symptoms were assessed using the CES-D and six DSM-oriented subscales of the Child Behavior Checklist (CBCL). Sleep parameters were objectively measured using wrist-worn actigraphy, from which sleep duration, wake after sleep onset (WASO), sleep efficiency, and Sleep Regularity Index (SRI) were extracted. Data analysis was conducted using R, including group comparisons, post-hoc analysis, and logistic regression. Results: The SR group demonstrated significantly lower SRI, higher WASO, and lower sleep efficiency compared to the other groups, along with elevated scores on the CES-D and CBCL subscales. Logistic regression identified lower SRI and higher somatic complaints as predictors of school refusal. Conclusions: Adolescents with school refusal exhibited marked irregularity in sleep–wake rhythms and impaired sleep quality. These disturbances were not fully explained by emotional symptoms. The Sleep Regularity Index (SRI) may serve as a behavioral marker for identifying school refusal. Future intervention models should integrate both emotional and chronobiological dimensions to improve treatment outcomes.