An app with no therapist attached can still treat insomnia

Cognitive behavioural therapy is the leading non-drug treatment for insomnia, but trained therapists are scarce and waiting lists are long, which is why fully automated digital versions have spread so quickly. These programmes deliver the entire therapy through software, with no clinician involved, making them scalable to millions of poor sleepers at once. Recent reviews have supported their overall efficacy, but uncertainty has lingered over whether the platform matters: do app-based programmes perform differently from web-based ones, or are apparent differences an artefact of how and when the trials were run? This review tackled both questions.

The authors conducted a systematic review and random-effects meta-analysis of randomised controlled trials of fully automated digital CBT for insomnia in general adult populations. Effects were pooled as standardised mean differences using restricted maximum likelihood estimation with a truncated Knapp-Hartung adjustment, a deliberately conservative approach. Subgroup analyses and an exploratory meta-regression then examined whether delivery platform and publication year moderated the results. Fifteen trials involving 3,507 participants met the inclusion criteria and contributed to the pooled analyses.

Fully automated digital CBT-I was associated with a substantial reduction in insomnia severity, at 0.82 standard deviations, alongside consistent improvements in sleep initiation and sleep efficiency. Improvements in overall sleep quality were smaller but still significant. In the exploratory multivariable meta-regression, the apparent association between delivery platform and treatment effect weakened after adjusting for study-level characteristics including publication year, and more recent trials tended to yield smaller effect estimates. Residual disagreement between the studies was low, though the small number of app-based trials means the platform analyses deserve caution.

The headline result confirms what earlier reviews have suggested: automated digital CBT-I works for adults with insomnia, and works well. The more novel contribution is the platform analysis, which complicates any claim that one delivery format is inherently better than another; differences across trials may reflect when and how studies were designed rather than the technology itself. The drift towards smaller effects in newer trials is a pattern worth watching, as it often appears when a field matures. Whether genuinely platform-specific effects exist awaits further trials, particularly app-based ones.

If insomnia has you weighing options and a therapist is out of reach, this is evidence that a fully self-guided digital programme can deliver a substantial reduction in insomnia severity without any human contact. It is also a reason not to fixate on whether a programme runs in a browser or an app, since platform differences did not hold up under scrutiny. The effect sizes in newer trials were somewhat smaller, so expectations should be realistic rather than miraculous. Persistent insomnia still warrants a proper clinical conversation.