Cognitive behavioural therapy for insomnia in people with cancer

Insomnia is a prevalent and distressing issue for individuals with cancer, negatively impacting their overall well-being. While cognitive behavioural therapy for insomnia (CBT-I) is the recommended first-line treatment for insomnia in the general population, its effects on cancer populations warrant rigorous, ongoing evaluation. Previous meta-analyses assessing CBT-I against specific comparators in people with cancer have often exhibited methodological limitations. Furthermore, with a recent increase in randomised controlled trials (RCTs) in this field, an updated and comprehensive synthesis is necessary.

To assess the effects of cognitive behavioural therapy for insomnia (CBT-I) in people with cancer.

In April 2025, we searched the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, Embase, Cumulative Index to Nursing and Allied Health Literature (CINAHL), PsycINFO, and clinical trials registries. We also checked reference lists. We applied no language, publication date, or study setting restrictions.

We included all RCTs that compared the effects of CBT-I with other treatments in individuals diagnosed with both insomnia and cancer. We placed no restrictions on the specific characteristics of the CBT-I interventions or the comparator groups. We excluded quasi-randomised and cross-over trials.

We used random-effects meta-analysis for our primary analyses. Our critical outcomes were: insomnia severity (Insomnia Severity Index (ISI), range 0 to 28), sleep quality (Pittsburgh Sleep Quality Index (PSQI), range 0 to 21), and serious adverse events (SAEs). Important outcomes were sleep diary-derived parameters: sleep onset latency (SOL), wake after sleep onset (WASO), total sleep time (TST), and sleep efficiency (SE). We also analysed objective sleep parameters for the comparisons presented in the summary of findings (SoF) tables. We used the Cochrane risk of bias 2 (RoB 2) tool to assess risk of bias and the GRADE approach to assess evidence certainty. We interpreted findings for continuous outcomes against minimally important differences (MIDs).

We included 21 RCTs (2431 randomised participants, predominantly female adults). Participants had received treatment for a range of cancer types, with breast cancer being the most frequently studied, and most were cancer survivors undergoing or following cancer treatment. CBT-I was typically delivered via therapist-led sessions or digital programs (four to 12 weeks). We identified five comparisons and summarised the findings for the two primary comparisons, with outcomes assessed at the conclusion of CBT-I treatment. The risk of bias was often high. Overall, CBT-I may offer small subjective improvements in insomnia for cancer survivors, and the evidence certainty was mostly low to very low. Benefits on objective measures are less clear. CBT-I versus inactive control Compared to inactive control post-intervention, CBT-I may slightly reduce ISI scores (mean difference (MD) -5.86 points, 95% confidence interval (CI) -7.22 to -4.51; 14 studies, 1371 participants; very low-certainty evidence) and PSQI scores (MD -3.60 points, 95% CI -4.95 to -2.24; 3 studies, 473 participants; low-certainty evidence). CBT-I may result in little to no difference in the occurrence of SAEs (risk ratio (RR) 1.05, 95% CI 0.07 to 16.77; 4 studies, 765 participants; very low-certainty evidence). Regarding sleep diary outcomes, CBT-I probably reduces SOL (MD -13.35 min, 95% CI -17.18 to -9.51; 9 studies, 760 participants; moderate-certainty evidence), may reduce WASO (MD -15.39 min, 95% CI -25.23 to -5.56; 9 studies, 784 participants; very low-certainty evidence), and may improve SE very slightly (MD 7.84%, 95% CI 3.62 to 12.06; 9 studies, 725 participants; very low-certainty evidence). CBT-I may result in little to no difference in TST (MD 6.43 min, 95% CI -8.30 to 21.16; 10 studies, 899 participants; very low-certainty evidence). Regarding objective sleep outcomes, CBT-I may result in little to no difference in SOL (MD -4.54 min, 95% CI -9.91 to 0.84; 2 studies, 129 participants), WASO (MD -5.08 min, 95% CI -10.62 to 0.46; 3 studies, 242 participants), TST (MD -11.01 min, 95% CI -29.35 to 7.33; 4 studies, 521 participants), and SE (MD 1.35%, 95% CI 0.08 to 2.63; 5 studies, 571 participants). CBT-I versus aerobic activities Compared to aerobic activities post-intervention, CBT-I may reduce ISI scores (MD -3.53 points, 95% CI -4.43 to -2.62; 2 studies, 406 participants; low-certainty evidence) and PSQI scores (MD -1.67 points, 95% CI -2.63 to -0.72; 3 studies, 496 participants; low-certainty evidence). CBT-I may result in little to no difference in the occurrence of SAEs (RR 0.35, 95% CI 0.04 to 3.36; 2 studies, 579 participants; very low-certainty evidence). Regarding sleep diary outcomes, CBT-I may result in little to no difference in SOL (MD -6.15 min, 95% CI -13.07 to 0.78; 2 studies, 131 participants; low-certainty evidence), WASO (MD -6.35 min, 95% CI -15.30 to 2.61; 2 studies, 131 participants; low-certainty evidence), and TST (MD 0.40 min, 95% CI -21.10 to 21.90; 2 studies, 131 participants; very low-certainty evidence). However, CBT-I may slightly increase SE (MD 4.24%, 95% CI 0.45 to 8.02; 2 studies, 131 participants; low-certainty evidence). Regarding objective sleep outcomes, CBT-I may result in little to no difference in SOL (MD 6.07 min, 95% CI -1.47 to 13.62; 2 studies, 131 participants), WASO (MD 6.16 min, 95% CI -3.80 to 16.11; 2 studies, 131 participants), and SE (MD -0.90%, 95% CI -2.84 to 1.04; 3 studies, 416 participants), and may lead to shorter TST (MD -13.02 min, 95% CI -25.00 to -1.04; 3 studies, 416 participants).

This Cochrane review found very low-certainty to moderate-certainty evidence suggesting that CBT-I may offer small to very small improvements in patient-reported insomnia severity, sleep quality, and most subjective sleep diary parameters when compared with inactive controls. Against aerobic activities, low-certainty evidence indicates that CBT-I may also improve insomnia severity and sleep quality. The incidence of SAEs appeared similar between CBT-I and comparator groups, and the certainty of this evidence is very low. The decision to use CBT-I to treat insomnia in people with cancer might depend on the treatment's availability and cost, as well as clinicians' and patients' preferences. Future research requires long-term, larger, and more rigorously designed studies that are inclusive and diverse.