Guided digital CBT for insomnia reduces sleep problems in adults with anxiety and PTSD symptoms

At the 2-month posttreatment assessment, CBT-I produced large reductions in insomnia severity (effect size d = -1.04; p < .001) compared to control, with these large effects persisting at 8-month follow-up (d = -0.98). Effects on anxiety symptoms and borderline personality symptoms were small (d = -0.28 and -0.36 respectively) at posttreatment and were no longer statistically significant at follow-up. Notably, CBT-I showed no significant effect on PTSD symptoms. Sleep diary measures showed small-to-moderate improvements, while improvements in broader positive health measures were minimal. From a cost perspective, CBT-I did not significantly reduce overall societal costs (difference of −€826), but demonstrated a 94% probability of cost-effectiveness at €450 per unit improvement in insomnia severity and 71% at the standard €20,000 per quality-adjusted life year threshold. These findings suggest CBT-I is an effective and cost-reasonable treatment specifically for insomnia in people with co-occurring psychological distress, though targeted interventions for anxiety, PTSD, or BPD may still be needed separately.

The trial enrolled 157 participants with insomnia and at least one co-occurring psychological condition, comparing 12 weeks of guided digital CBT-I against a waitlist control receiving only sleep diaries. The insomnia improvements were robust—the effect size of d = −1.04 at posttreatment is considered large in psychological research and remained sustained 6 months later. However, the secondary psychological effects tell a more nuanced story: anxiety and BPD symptom reductions (d = −0.28 to −0.36) were small in magnitude and did not persist at follow-up, meaning participants' underlying anxiety did not durably improve. PTSD showed no meaningful change at any timepoint. Sleep diary data (nights awake, time in bed, sleep efficiency) improved modestly, suggesting participants slept somewhat better objectively, though subjective wellbeing measures did not shift meaningfully. The cost-effectiveness analysis, rare in psychotherapy trials, found CBT-I did not save money overall but offered reasonable value—a 71% probability of meeting societal willingness-to-pay thresholds—because insomnia relief itself carries meaningful quality-of-life benefit. The main limitation is that this was a single-arm comparison against a minimal control, so some improvement may reflect placebo or natural recovery rather than the specific therapy.