Determination of Optimal Sleep Treatment Elements - MOST

The DOSE Project employs the Multiphase Optimization Strategy (MOST) to evaluate the relative efficacy of the individual cognitive behavioural therapy for insomnia (CBT-I) components. The present study represents the optimization phase of MOST, which aims to identify which combination of CBT-I components produces the best expected outcome, while taking constraints, such as economy, scalability, treatment complexity, and usability, into account. The main objectives of the DOSE Project are 1. To conduct MOST on a multi-component smartphone or web application targeting insomnia 2. To identify the components and combination of components that demonstrate the best expected obtainable outcome 3. To examine potential moderators of the effect of the individual treatment components on insomnia severity, 4. To determine which application build (i.e., combination of components) is optimal to test in a subsequent RCT. This study employs a fractional factorial experimental design conducted with individuals experiencing moderate-to-severe insomnia. With random allocation, participants will receive a version of the application in which one or more of the five core CBT-I components (i.e., sleep hygiene education, sleep restriction, stimulus control therapy, deactivation/relaxation training, cognitive therapy) are presented in various combinations. The intervention lasts 10 weeks, including an initial one-week assessment period. Baseline group differences (concerning socio-demographic, disease-related, and psychosocial data) will be explored to test the success of the randomization. If differences are found, sensitivity analyses will be made to evaluate their possible influence on the results. Main effects will be analysed using Mixed Linear Models (MLMs) based on the intent-to-treat sample, comparing aggregated groups of N = 2 x 320 (e.g., plus/minus sleep restriction or plus/minus sleep hygiene etc.) on all outcome variables. MLMs account for the hierarchical, non-independent nature of the data (i.e., repeated measures nested within patients and treatment conditions), testing the time\*group interaction effect, reflecting the effect of treatment. Moderation analyses will evaluate whether individual differences in various baseline variables (e.g., physical function, expectations, computer proficiency, chronotype, etc.) influence intervention effects.