緒論 失眠困擾對身心健康有重大的影響，除了藥物取向的治療外，失眠認知行為治療（CBT-I）有足夠的實證研究證據肯定其療效，由於在認知行為治療（CBT）的研究中有證據支持參與者在家中，對治療技術練習與執行的程度與療效有關，因此CBT-I的療效可能也與患者對治療技術的執行程度有關，過去雖然也有少數研究探討CBT-I執行程度與療效的關係，但這些研究缺乏針對CBT-I各個治療技術的執行程度評估，因此本研究欲探討CBT-I中，各項治療技術（睡眠衛生教育、放鬆訓練、刺激控制法、睡眠限制法，及認知重建）的執行程度及執行規律程度，與特定療效指標改善程度的關係。 方法 研究參與者含原發性失眠患者22位，共病其它疾患的失眠患者16位，經篩選階段確認符合收案條件後，在接受為期七週的CBT-I團體治療期間，於第二週起需每日填寫執行程度問卷，同時帶領團體的治療師也於治療第三週起，以治療師執行程度評估量表來評估參與者的執行程度。參與者在治療前後另需分別填寫一週睡眠日誌以及睡眠困擾問卷（Insomnia Severity Inventory；ISI），以睡眠日誌中的SOL、WASO、TST、SE，及ISI得分的前後測改變分數，作為代表療效指標之依變項，預測變項則為參與者自評及治療師所評估的各項治療技術之執行程度分數，治療技術包括：睡眠衛生、放鬆訓練、刺激控制法、睡眠限制法，以及認知重建；並分別以平均數代表執行程度，而以變異數代表執行期間的規律程度。 結果 資料分析以皮爾森相關分析檢驗執行程度與療效間的關聯性，在全部樣本中的結果發現，參與者自評對認知重建的平均數與WASO的改善程度有正相關，以及治療師評估CBT-I的平均數與SOL的改善程度有正相關；而參與者自評放鬆訓練的平均數越低、刺激控制法變異數越高，則ISI的改善程度越好，為不符合預期的結果。在原發性失眠組中發現治療師評估睡眠衛生的平均數與WASO的改善有正相關，治療師評估睡眠限制法的平均數與SOL、TST，以及SE的改善有正相關，而變異數與SOL的改善有負相關，以及治療師評估CBT-I的平均數與SOL的改善有正相關；而不符合預期的結果為，參與者自評執行刺激控制法的變異數與ISI的改善有正相關。最後，在共病組中並沒有發現執行程度與療效之間有顯著相關的結果，而放鬆訓練以及刺激控制法兩項治療技術，也沒有與療效指標有符合預期的顯著相關結果。 結論 對於共病失眠患者而言，治療技術的執行評估與療效間未反映出顯著相關，可能因受限於睡眠生理疾患的干擾，使得療效未如原發性失眠組明顯；而就原發性失眠組而言，睡眠限制法的執行程度是與較多療效指標達到顯著相關的治療技術，顯示睡眠限制法的執行對於改善睡眠症狀有所幫助。本研究較為重大的限制在於，治療技術執行程度評估的評分者間一致性有限，後續研究可進一步發展評估執行程度更適當的方法。 / Introduction Insomnia is a common problem that has a significant impact on patients' physical and mental health. In addition to pharmacological therapy, there are sufficient empirical data to support treatment efficacy of cognitive behavioral therapy for insomnia (CBT-I). A major part of cognitive behavioral therapy is to teach the patients to learn specific cognitive and behavioral techniques that requires to be practiced at home by the patients. Previous studies have reported an association between the degree of adherence to treatment techniques and treatment effects in cognitive behavioral therapy for disorders other than insomnia. Some studies further explored the relationship between treatment adherence and treatment outcome in CBT-I. However, none of the study looked into the adherence to different treatment components on different aspects of treatment outcome. The aim of this study therefore is to explore the relationship between the adherence to different treatment components in CBT-I and the improvement in different sleep parameters. Method Twenty-two patients with primary insomnia and sixteen patients with comorbid insomnia were recruited for this study. During the 6-week period of treatment with CBT-I, they completed a treatment adherence questionnaire daily. The therapists also evaluated the patients’ adherence weekly after treatment sessions. They were required to keep sleep diaries from one week before to one week after the end of the 6-week CBT-I program. They also completed the ISI and sleep diary for one week before and after the treatment. The treatment outcome variables included the ISI score, and sleep onset latency (SOL), wake after sleep onset (WASO), total sleep time (TST) and the sleep efficiency (SE) from sleep diaries. The treatment components evaluated included sleep hygiene, relaxation, stimulus control, sleep restriction and cognitive reconstruction. The predictive variables included the mean for the adherence degrees and the variance for the adherence regularity of the adherence score evaluated by the participants and therapists. Results Pearson correlation was used to examine the associations between the adherence to respective treatment components in CBT-I and the variables of treatment outcome. In the whole sample, the decrease of the WASO correlated significantly with adherence to cognitive reconstruction evaluated by the patients, and the decrease of the SOL correlated significantly with adherence to CBT-I evaluated by the therapists. Nonetheless, the lower adherence to relaxation and the higher variance of adherence to stimulus control, the more improvement of the ISI. In patients with primary insomnia, the decrease of WASO correlated significantly with adherence to sleep hygiene evaluated by the therapists; the decrease of SOL and the increase of TST and SE showed significant correlation with adherence to sleep restriction evaluated by the therapists; the decrease of SOL showed significant negative correlation with the variance of adherence to sleep restriction evaluated by the therapists; the decrease of SOL showed significant positive correlation with the adherence to CBT-I evaluated by the therapists. Opposite to expectation, the decrease of the ISI score correlated significantly with variability of adherence to stimulus control evaluated by the patients. Finally, in patients with comorbid insomnia, the treatment component adherence did not correlate significantly with any outcome variables. The adherence of relaxation and stimulus control did not show significant correlation with outcome variables. Conclusion In patients with comorbid insomnia, treatment effects were not associated with adherence. This may due to the multifactorial nature of comorbid insomnia. Among the treatment components, adherence to sleep restriction seems to be the most predictive factor for good treatment outcome in primary insomnia. One limitation of the present study is its low inter-rater reliability of treatment components adherence evaluation. Therefore, more appropriate methods need to be developed to evaluate the adherence of treatment techniques.