Dismantling cognitive-behavioural therapy for chronic insomnia in adults with or without comorbidities: a systematic review and component network meta-analysis

BackgroundChronic insomnia disorder is highly prevalent, disabling and costly. Cognitive-behavioural therapy for insomnia (CBT-I) is recommended as the first-line treatment. CBT-I may contain various educational, cognitive and behavioural strategies delivered in a range of formats, but the effect of each component remains unclear. MethodsWe performed a systematic review and component network meta-analysis (cNMA) of CBT-I trials for chronic insomnia. We searched PubMed, CENTRAL, PsycINFO and ICTRP for randomised controlled trials published from database inception to 14th May 2022, comparing any form of CBT-I against each other or a control condition for chronic insomnia disorder in adults (aged [≥]18 years). We included insomnia both with and without comorbidities. Concomitant treatments were allowed if they were equally distributed among the arms. Two independent reviewers identified components, extracted data, and assessed trial quality. Primary outcome of interest in this study was treatment efficacy (remission defined as reaching a satisfactory state at endpoint measured by any validated self-reported scale) at post-treatment. (PROSPERO; CRD42022324233) FindingsWe identified 226 trials, including 29,982 participants. Mean age was 45{middle dot}7 years and 71% were women. The results suggests that critical components of CBT-I are cognitive restructuring (incremental odds ratio[iOR] 1{middle dot}63 [95% confidence interval 1{middle dot}25 to 2{middle dot}14]), sleep restriction (iOR 1{middle dot}44 [1{middle dot}00 to 2{middle dot}06]) and stimulus control (iOR 1{middle dot}44 [1{middle dot}00 to 2{middle dot}07]) Sleep hygiene education was not essential (iOR 1{middle dot}05 [0{middle dot}79 to 1{middle dot}38]) and relaxation procedures may be counterproductive (iOR 0{middle dot}81 [0{middle dot}64 to 1{middle dot}03]). Face-to-face, therapist-led program was found to be most beneficial (iOR 1{middle dot}86 [1{middle dot}21to 2{middle dot}85]). The overall risk of bias was low in 8% of the trials, some concerns in 56%, and high in 36%. InterpretationThis cNMA suggests that effective and efficient CBT-I packages can include cognitive restructuring, sleep restriction and stimulus control, but not relaxation. FundingNone.