Investigators tested i-Sleep Autism, an autism-adapted, guided, internet-delivered cognitive behavioral therapy for insomnia, against a waitlist control that received psychoeducation and sleep hygiene. The intervention spanned five weeks and was co-created with seven autistic adults; the source reported that i-Sleep Autism was adapted from a pre-existing non-autism-specific i-Sleep program. A total of 163 adults meeting the Insomnia Severity Index (ISI) threshold of ≥ 10 were randomized: 82 to the intervention and 81 to the control. Assessments were scheduled at baseline (T0), mid-intervention (T1; three weeks), post-test (T2; six weeks), and follow-up (T3; six months). The primary endpoint was insomnia severity. The source did not report additional trial details such as recruitment settings or specific inclusion/exclusion criteria beyond ISI ≥ 10.
Adherence varied. The study team observed that 56.1% of participants in the intervention group completed four modules and 50% completed all five modules. The report did not provide granular week-by-week adherence curves or reasons for discontinuation. Engagement therefore appears moderate, with roughly half the intervention arm completing the full five-module program; the authors signaled that dropout and variable engagement remain implementation challenges that require targeted strategies.
Intent-to-treat analyses indicated a statistically and clinically meaningful reduction in insomnia severity for the i-Sleep Autism group compared with waitlist at mid-intervention and at post-test. Data indicated an effect size of d = 1.02 with a 95% confidence interval of [0.65, 1.40] when comparing intervention to control at post-test. The source stated that these improvements were significant at both T1 and T2, documenting a large magnitude of benefit on the trial’s primary outcome. The article did not report additional statistical details such as p-values or model covariates in the summary abstract, so those elements are not available from the source text.
The study team reported reductions in both cognitive and somatic pre-sleep arousal at post-test in the intervention arm versus control. Quality of life scores were higher at post-test for the intervention group. By contrast, the investigators did not find significant effects on anxiety, depression, or daily functioning at the assessed time points. These null results suggest that while targeted digital CBT-I reduced sleep symptoms and some associated arousal domains, it did not translate into measurable improvements in broader mood symptoms or functional outcomes within the trial’s timeframe.
The authors observed that effects on insomnia severity were maintained at the six-month follow-up (T3). At post-test, 37.8% of participants in the intervention arm met criteria for clinically meaningful improvement in insomnia compared with 11.1% in the control arm. The source reported maintenance of effects at T3 but did not provide a detailed breakdown of responder status at six months in the abstract. Those seeking full responder trajectories should consult the full text for itemized follow-up data and responder analyses.
The investigators noted considerable variation in engagement and in individual response to the intervention. They recommended further work to identify who benefits most, to optimize support for non-responders, and to improve accessibility and adherence. The source did not report specific alternative strategies or adjunctive approaches to address dropouts. The abstract also did not include detailed demographic breakdowns, comprehensive adverse event reporting, or trial operational details such as blinding procedures—information that readers will need to extract from the full article. These gaps represent the primary unanswered clinical and implementation questions documented in the source.