The comparative efficacy of various approaches of digital cognitive behavioral therapy for insomnia (CBTi) is still unclear. This network meta-analysis explored the comparative efficacy of digital CBTi approaches in adults with insomnia. Four electronic databases were searched from inception to June 27, 2020. Primary outcomes were self-reported total sleep time (TST), sleep onset latency (SOL), wake after sleep onset (WASO), sleep efficiency (SE), and insomnia symptoms; these were measured using sleep diaries or valid questionnaires. A random-effects network meta-analysis in a frequentist framework was used. Fifty-four randomized controlled trials comprising 11,815 participants were included. Compared with usual care, web-based CBTi with a therapist demonstrated significantly longer TST (mean difference [MD]: 23.19 min, 95% confidence interval [CI]: 18.98e27.39 min), shorter SOL (MD:
18.76 min, 95% CI
24.20 to
13.31 min), lower WASO (MD:
31.40 min, 95% CI:
36.26 to
26.55 min), and greater SE (MD: 10.37%, 95% CI: 8.08%e12.65%). The surface under the cumulative ranking curve indicates that webbased CBTi with therapists is most likely to be ranked the highest among all treatments, and thus, this network meta-analysis suggests that such a treatment is the optimal intervention for improving sleep duration and SE as well as the reductions in SOL and WASO. Prospero registration number: CRD42020171134.