RESUMO
It is well-established that cognitive behavioural therapy for insomnia (CBT-I) improves self-reported sleep disturbance, however the impact on objective sleep is less clear. This meta-analysis aimed to quantify the impact of multi-component CBT-I on objective measures of sleep, indexed via polysomnography (PSG) and actigraphy. Fifteen studies met inclusion criteria. Following appraisal for risk of bias, extracted data were meta-analysed using random-effects models. The quality of the literature was generally high, although reporting of methodological detail varied markedly between studies. Meta-analyses found no evidence that CBT-I reliably improves PSG-defined sleep parameters. Actigraphy evidence was more mixed; with a small effect for reduction in sleep onset latency (Hedge's g = -0.28 [95% confidence interval (CI) -0.51 to -0.05], p = 0.018) and a moderate effect for reduction in total sleep time (TST) (Hedge's g = -0.51 [95% CI -0.75 to -0.26], p < 0.001). In contrast, and consistent with recent meta-analyses, CBT-I was associated with robust improvements in diary measures of sleep initiation and maintenance (Hedge's g range = 0.50 to 0.79) but not TST. While the literature is small and still developing, the sleep benefits of CBT-I are more clearly expressed in the subjective versus objective domain.
