0432 Baseline Insomnia Severity Predicts Magnitude of Response to Digital Cognitive Behavioral Therapy for Insomnia: Greater Improvements in Severe Cases

Abstract Introduction Identifying baseline predictors of treatment response is essential for patient stratification and clinical expectation management in insomnia interventions. While evidence indicates that severe insomnia may demonstrate reduced response to behavioral treatments, research examining digital therapeutic platforms remains limited. This study evaluated whether baseline insomnia severity, quantified by sleep efficiency, predicts treatment response to digital cognitive behavioral therapy for insomnia (dCBT-I). Methods Adults (N=528) completed an 8-week digital CBT-I program with baseline and post-intervention sleep diaries. Participants were stratified by baseline sleep efficiency: Severe (< 70%, n=108), Moderate (70-84%, n=197), and Mild (≥85%, n=212). Primary outcomes were absolute and relative changes in sleep efficiency. Secondary outcomes included total sleep time, sleep onset latency, and awakenings. Linear regression examined baseline sleep efficiency as a predictor of treatment response, controlling for age and gender. Results Baseline severity strongly predicted treatment response magnitude. Severe insomnia participants showed the largest improvements: sleep efficiency increased from 55.0±11.2% to 79.2±13.8% (absolute change +24.2 percentage points, relative improvement +44.0%, p< 0.001, Cohen's d=1.52). Moderate insomnia participants improved from 78.7±4.1% to 86.8±10.5% (+8.1 percentage points, +10.3%, p< 0.001, d=0.51). Mild insomnia participants showed minimal change from 90.1±4.2% to 91.2±7.8% (+1.1 percentage points, +1.2%, p< 0.05, d=0.09). Linear regression confirmed baseline sleep efficiency as a significant predictor of absolute improvement (β=-0.48, p< 0.001), with each 10-point lower baseline sleep efficiency predicting an additional 4.8 percentage points improvement. Despite starting with lower sleep efficiency, severe insomnia participants reached clinically meaningful levels (79.2%) by week 8. Conclusion Digital CBT-I demonstrates greatest efficacy in individuals with severe baseline insomnia, with effect sizes (d=1.52) exceeding those typically reported for behavioral interventions. These findings challenge concerns about treatment effectiveness in severe cases and suggest digital CBT-I should be prioritized for patients with marked sleep disturbance where potential for meaningful improvement is greatest. The strong dose-response relationship between baseline severity and treatment response provides evidence-based expectations for outcomes. Support (if any)