Effects of cognitive-behavioral therapy for insomnia compared with controls among cancer survivors: a systematic review and meta-analysis of randomized trials.

backgroundInsomnia is highly prevalent among cancer survivors. Meta-analyses examining the effects of cognitive-behavioral therapy for insomnia (CBT-I) among cancer survivors have focused on within-group (pre-to-post-intervention) changes, with calls to better evaluate treatment effects.

objectiveTo conduct a systematic-review and meta-analysis and evaluate the effects of cognitive-behavioral therapy for insomnia (CBT-I) among cancer survivors, compared with controls, on insomnia.

methodsWe followed recommendations from the Cochrane Handbook and PRISMA guidelines. We comprehensively searched 8 databases (CINAHL/ClinicalTrials.gov/Cochrane Central/Embase/MEDLINE/PEDro/PsychInfo/Web of Science) and included randomized controlled trials (RCTs) in which adult cancer survivors with clinically-significant insomnia were randomized to CBT-I or control conditions that included usual care, wait-list, attention, or sleep hygiene education only. We designated the primary outcome as end-of-intervention Insomnia Severity Index (ISI) and secondary outcomes included sleep diary parameters, fatigue, and health-related quality of life (HRQL). We analyzed between-group mean differences (MD's), standardized-mean-differences (SMD's), and interpreted results using minimal clinically important difference (MCID) thresholds as endorsed by the American College of Physicians (ACP) or SMD thresholds. We rated evidence certainty using GRADE, facilitated by GRADEpro GDT.

resultsWe included 19 RCTs involving 1,803 participants. Participant mean age was 55 and time-since-diagnosis was 2.5 years; 94% were women, mostly survivors of breast cancer. At end-of-intervention, compared with controls, CBT-I improved ISI [MD (95% CI): -4.4 (-5.3, -3.5) points; assessed in 13 trials] that did not reach the MCID threshold (i.e., ≥ 6 points) to suggest that many patients derived clinically-important benefit, but is higher than half of the minimal-important-change (MIC) (i.e., 3-<6 points, including 95% CI), suggesting that an appreciable number of patients derived clinically-important benefit. Subjective sleep diary (assessed in 12 trials) sleep latency, wake after sleep onset, sleep efficiency, fatigue (11 trials), and HRQL (10 trials) were also improved; however, on average, none of the improvements reached their respective MCID or SMD thresholds to suggest that many patients derived clinically-important benefits. In pre-specified subgroup analyses, no intervention or cancer-related characteristics meaningfully changed results. Evidence certainty was low-to-very-low, primarily due to heterogeneity, performance, publication, and/or reporting bias.

conclusionCompared with controls, CBT-I improved insomnia at an average magnitude greater than half of the MIC but did not reach the MCID threshold, suggesting that an appreciable number, not many, of cancer survivors derived clinically-important benefit. Strategies are needed to improve insomnia for many cancer survivors, particularly among non-responders to first-line CBT-I. PROTOCOL REGISTRATION: PROSPERO (CRD42022332584).