Digital CBT-I for Patients With Chronic Pain and Insomnia

Overview

Chronic pain constitutes an increasing health and social burden. More than 50% of patients with chronic pain report insomnia, and patients with comorbid insomnia often report stronger and widespread pain, compared to those who are sleeping well. Sleep disturbances are often considered a consequence to chronic pain. This means that insomnia is often overlooked or ineffectively managed with hypnotics or advice on sleep hygiene. Therefore, efficacious, easily accessible, and safe alternatives to the current pharmacological treatments for patients with chronic pain and insomnia are needed. Cognitive behavioral therapy for insomnia (CBT-I) is a cost-effective and safe treatment for insomnia and is recommended as first-line treatment. While highly efficacious, the challenge is to deliver CBT-I to those in need. The main barriers of face-to-face delivered CBT-I are availability of trained therapists, costs, as well as physical and geographical constraints. The primary aim of this randomized controlled clinical superiority trial is to investigate whether digitally-delivered CBT-I has a greater effect on insomnia and pain than digitally-delivered sleep hygiene education in patients with chronic pain and comorbid insomnia. Secondary objectives are to a) explore whether the pain-relieving effect is mediated by a change in physiological markers of sleep quality, b) whether health care cost and use of medications at 12 months are reduced after digital CBT-I, and c) to explore the effectiveness of digital CBT-I compared with sleep hygiene education on: 1. Physiological sleep metrics (recorded with ear EEG in subsample of 60 patients). 2. Self-reported sleep quality. 3. Quality of life. 4. Physical and mental health. 5. Thoughts and beliefs about sleep and pain.

This study aims to investigate the efficacy of digital CBT-I for improving sleep (i.e. reducing insomnia severity) and reducing pain intensity in patients with chronic pain and comorbid insomnia. To achieve this goal, the investigators will perform a randomized controlled single-blind clinical superiority trial. Participants will be randomized with an allocation 1:1 ratio into a digital CBT-I group and a sleep hygiene education group. Both groups will use the Hvil® app, but the amount of content will be different. Using a data manager not affiliated with the project through the Open Patients data Exploratory Network (OPEN) at Odense University Hospital, the randomization will be implemented in the program REDCap. The data manager will prepare the randomization list in REDCap. The computer algorithm will use permuted-block randomization of 2, 4 and 6 individuals. No stratifications will be applied to the randomization. After inclusion, participants will receive an unique randomization key to impute in the Hvil® app according to the abovementioned procedure to receive either digital CBT-I or sleep hygiene education for 9 weeks. This Hvil® App platform is used for both treatments to ensure that only the contents is different, but the form of delivery is the same. All patient reported data will be filled in directly to the eCRF via an online questionnaire using REDCap. Statistical analyses will be blinded to treatment allocation. Sample size estimation: Minimal clinical important difference for insomnia severity is 5 point. With an estimated SD for change of 2 times the minimal clinical important difference (SD=10) for patients with chronic pain and insomnia, power of 80%, an alpha value of 0.05, a minimum of 63 participants in each group is required. Minimal clinical important difference for pain intensity is 15% corresponding to a change of 1.0 points in this population. With a SD for change of 2.0 for patients with chronic pain and insomnia, power of 80%, an alpha value of 0.05, a minimum of 63 participants in each group is required. With an expected drop-out of 30% the investigators plan to recruit 160 participants (80 for each group). Stopping rules: This trial has two stopping rules. Inclusion will stop December 31st 2025. In the event that 130 patients have completed the 9 weeks follow-up before 160 patients have been recruited the investigators will stop recruitment of further participants. All participants that are withdrawn from the study will be encouraged to complete all visits as scheduled. Participants that are withdrawn from the treatment will not be replaced and will be included in the intention-to-treat population. According to the sample size calculation, the investigators will include participants enough to tolerate a drop-out rate of up to 30%. Both intention-to-treat and per protocol analysis will be performed and compared to assess the robustness of the primary analysis. Statistical methods: Difference in change from baseline to 9 weeks in the two primary outcomes (ISI score or mean NRS pain in last 7 days, both being continuous variables) between the digital CBT-I group and the sleep hygiene education group will be estimated using a mixed linear model approach with ISI and NRS pain as outcomes, time, treatment arm (digital CBT-I or sleep hygiene education), and the interaction between time and treatment arm as fixed effects. Both participant specific intercept and slope will be considered as random effects. The model will be adjusted for ISI score or NRS pain at baseline, respectively. The error variance will be allowed to vary over time and across the two arms. Interactions will be parametrized as time dependent treatment effects such that the treatment effect at week 9 corresponds directly with the efficacy parameter of interest. The treatment effect will be expressed as the gain in reduction observed in the digital CBT-I arm, i.e. a positive number expresses a favoring of the intervention. The pre-specified effectiveness analyses will be based on the intention-to-treat (ITT) principle, which includes all participants that are assessed and randomized at baseline. In the case of missing data during the 9-week trial, repeated measures linear mixed models will take this into account automatically, under the assumption that data is missing at random. For the statistical assessment of the two primary outcomes (ISI and pain intensity), the investigators will apply a sequential test procedure to compare the effect of the two treatments (CBT-I vs. sleep hygiene education) on ISI severity in the first step and the effect on pain intensity in the second step. This means that statistical significance in the second step will only be assessed if statistical significance has been reached in the first step . The investigators use this approach as it seems unlikely that these interventions has direct effect on pain, but rather work through improved sleep. The significance level is set to 5%. In responder analysis, a responder is defined as a patient who report a more than 30% and 50% decrease in insomnia or pain after 9 weeks. Proportion of responders between treatment groups will be expressed by adjusted odds ratios. These will be based on a logistic regression model with treatment, baseline value of pain intensity and insomnia severity, and sex as covariates. A detailed statistical analysis plan (SAP) will be made publicly available before the statistical analysis is initiated.