The Effect of Digital Cognitive Behavioural Therapy for Insomnia on Physical Activity in Fibromyalgia

Overview

The goal of this clinical trial is to learn if a digital sleep therapy program (digital Cognitive Behavioural Therapy for Insomnia or dCBT-I) works to improve quality of life and movement in adults with fibromyalgia who also have trouble sleeping. The main questions it aims to answer are: * Does digital sleep therapy improve quality of life for people with fibromyalgia? * Does digital sleep therapy improve sleep quality? * Does better sleep help reduce fear of movement and increase physical activity, assessed in a virtual reality (VR) environment? Researchers will compare digital sleep therapy (called 'Sleepio') to standard care with sleep advice materials. Participants will: * Use the Sleepio program at home for 10 weeks (6 sessions, 20 minutes each) * Wear a sleep monitoring device at home in bed for several nights at the start and after 3 months * Wear an activity watch for 1 week to track movement at the start and after 3 months * Complete questionnaires about pain, sleep, mood, and daily activities at the start, 3 months, and 6 months * Visit the study centre twice for assessments that include: * Pain sensitivity testing * A virtual reality game that measures how they move and make decisions * Recording of simple exercises like marching and squats Participation in the study lasts about 6 months.

BACKGROUND AND RATIONALE Pain and sleep share a complex bidirectional relationship. In both laboratory and clinical settings, pain demonstrates circadian rhythmicity and sensitivity to sleep deprivation. Poor sleep aggravates pain, and increasing pain worsens sleep disturbance, creating a potentially self-perpetuating cycle that may contribute to the development and maintenance of chronic pain. However, the mechanisms by which sleep deprivation increases pain remain unclear, particularly in the context of fibromyalgia. One proposed mechanism involves kinesiophobia (fear of movement), which is central to the Fear Avoidance Model of chronic pain. This model proposes that excessive fear of movement, due to the expectation that moving will cause pain, prevents people from engaging in beneficial physical activities. This avoidance leads to physical deconditioning, which further amplifies pain by reducing the engagement of endogenous descending pain control mechanisms. Understanding the mechanistic link between sleep, pain, and physical activity is particularly important in fibromyalgia, a debilitating chronic pain condition characterised by widespread pain, poor sleep, fatigue, and reduced motivation. Polysomnography studies in fibromyalgia have demonstrated specific EEG correlates of sleep fragmentation, including alpha intrusion into slow-wave sleep, which may reflect non-restorative sleep. Observational and laboratory studies have shown behavioural changes consistent with kinesiophobia in this population. Previous research by our group has found that 73% of fibromyalgia patients fulfil criteria for clinical insomnia, with strong correlations between sleep disturbance and pain severity (R=0.44; P\<0.001), fear of movement (R=0.38; P\<0.001), and quality of life measures (R=0.51; P\<0.001). Additionally, patient partners have explicitly identified poor sleep as a barrier to participating in physical rehabilitation and regular exercise, both of which are mainstays of fibromyalgia treatment. RATIONALE FOR DIGITAL CBT-I Recent systematic reviews suggest that cognitive behavioural therapy for insomnia (CBT-I) is associated with significant improvements in self-reported sleep quality, pain, and depression in chronic pain populations. However, existing research has important limitations: (1) it has not focused on physiological measures of sleep and activity, (2) quantitative measures of motor performance and pain have not been assessed, (3) the mechanisms by which CBT-I affects pain and physical activity remain unknown, and (4) most studies have used in-person CBT-I, which has variable delivery and limited accessibility. Digital CBT-I (dCBT-I) addresses these limitations. It removes inter-therapist variability, ensuring consistent treatment delivery, and overcomes access barriers such as therapist shortages, long waiting lists, and travel difficulties. Sleepio, the dCBT-I program used in this study, has demonstrated effectiveness for sleep disturbance and cognitive symptoms in insomnia and has NICE accreditation. STUDY DESIGN AND INTERVENTION This is a randomised controlled trial (RCT) evaluating the efficacy of dCBT-I in improving fibromyalgia-related quality of life, pain, sleep physiology, and physical activity. The study aims to understand the mechanisms through which sleep affects pain and movement in fibromyalgia. Participants will be randomised 1:1 to either Sleepio or standard care after a 4-week run-in period. Intervention Group: Participants will receive access to Sleepio, which delivers 6 sessions of automated CBT-I over 10 weeks (approximately 20 minutes per session). The program features an animated virtual professor and includes evidence-based cognitive and behavioural interventions, sleep hygiene education, time-in-bed restriction, and relaxation exercises. Participants also have access to a daily online sleep diary, an online community, and a resource library. The program can be accessed via web browser or smartphone app at the participant's convenience. To support adherence, text reminders will be sent through REDCap, and participants will be contacted by a study investigator via telephone or video call at 1, 3, and 6 weeks. Control Group: Participants receive standard NHS care plus written materials from Versus Arthritis providing evidence-based advice on sleep hygiene. To standardize care, the intervention group also receives access to these same sleep hygiene materials. NOVEL ASSESSMENT METHODOLOGY Understanding how sleep affects pain and movement requires a multi-modal measurement approach. This study employs a comprehensive assessment battery that combines subjective reports with objective physiological and behavioural measures. Sleep Assessment Home Sleep EEG: Physiological sleep is measured using a self-applied, single-channel EEG device (SOMNOmedics Home Sleep Test HST-REM system) coupled with electrooculography (EOG), electromyography (EMG), and sensors for ambient light, activity, body position, and snoring detection. Recordings will be collected for a minimum of 4 days prior to randomisation and after treatment at 3 months. This allows assessment of sleep regularity, sleep continuity (sleep onset latency, wake after sleep onset, sleep efficiency, total sleep time), and sleep architecture (arousal indices, time in sleep stages). Actigraphy: Participants will wear a research-grade actigraphy device (CamnTech MotionWatch 8) continuously for 7 days to measure rest/activity patterns, overall physical activity levels, and sleep/wake patterns in the home environment. Under-mattress Sensor: The Withings Sleep Analyzer will be placed under the participants' home mattresses for the duration of the intervention period to continuously measure movement, heart rate, respiratory rate, body position, and snoring, providing additional data on sleep timing and estimated sleep stages. Pain Assessment Quantitative Sensory Testing (QST): Pain sensitivity will be assessed using a standardised QST protocol developed by Rolke and the German Research Group on Neuropathic Pain. Testing measures skin sensitivity to thermal, touch, and vibration stimuli. Assessments are conducted over areas primarily affected by fibromyalgia as well as control areas not directly involved. Key parameters include mechanical pain threshold, wind-up ratio, and pressure pain thresholds. This provides objective, quantifiable measures of pain processing and central sensitization. Electrical Pain Stimuli: Brief, mildly painful electrical stimuli are delivered during the virtual reality task using a Digitimer DS7A constant current stimulator. Pain intensity is individually calibrated prior to tasks using a standard staircase procedure, with participants' pain threshold identified and stimulus intensity set to achieve a moderately painful sensation (6/10 on a Likert scale). Movement Assessment The study will employ three complementary approaches to assess movement behaviour, representing a novel multi-modal evaluation of kinesiophobia and physical function: 1. Virtual Reality (VR) Task: This laboratory-based assessment measures kinesiophobia during ecologically realistic behaviour. Participants use an HTC Vive headset to search a virtual jungle for fruit scattered on the floor. Fruits win game points, but some (distinguishable by colour) are "spiky" and trigger a brief electrical stimulus when handled. The VR environment allows precise measurement of movement patterns during motivated behaviour in ways not possible with conventional computerized tasks. Computational deconstruction of behaviour through statistical model-fitting yields summary parameters including: horizontal movement tendency and vertical movement tendency (the primary measures of kinesiophobia), motivational vigour (a measure of fatigue), reward sensitivity (related to mood), and pain sensitivity (related to anxiety). This paradigm was co-designed with patient partners, and pilot data validated using an experimental tonic pain model demonstrated substantial reductions in horizontal and vertical movement correlated with pain intensity. 2. Quantitative Movement Testing (QMT): Video recordings capture participants performing three standardized physiotherapy-type movements (marching on the spot, mini squats, and forward bends) using a smartphone or tablet camera. Motion capture analysis employs university-owned software based on Detectron2, utilizing a pre-trained dilated temporal convolutional neural network (VideoPose3D) for pose extraction. This approach captures long-term temporal information with higher accuracy, simplicity, and efficiency compared to traditional recurrent neural network models. The system extracts 2D/3D skeletal poses and computes movement metrics such as joint angles using linear algebra operations applied to filtered 3D pose data. This provides quantifiable measures of physical ability and pain-related changes in movement. Three repetitions of each movement are recorded to assess consistency and variability. The original videos are not stored; only the computer-generated skeletal coordinate time series are retained. 3. Actigraphy-Based Physical Activity: Home-based activity monitoring provides naturalistic data on overall physical activity levels, activity patterns throughout the day, and sedentary behaviour in the participants' usual environment. STUDY ASSESSMENTS AND TIMELINE Participants will be followed for approximately 6 months (29 weeks) with assessments at baseline (prior to randomisation), 3 months, and 6 months. The baseline and 3-month visits include in-person assessments (approximately 2 hours each) consisting of questionnaires, QST, VR tasks, and motor evaluation, along with home-based sleep monitoring and actigraphy. The 6-month assessment consists of online questionnaires only (approximately 30 minutes). Questionnaires will be completed online via REDCap. The 4-week run-in period before randomisation allows baseline data collection and ensures participants meet study criteria. For the intervention group, the 10-week Sleepio program begins after randomisation, with the 3-month assessment timed to capture immediate post-intervention effects. BLINDING Researchers analysing physiological data (sleep EEG, actigraphy) and conducting statistical analyses remain blind to treatment allocation. While participants and study coordinators cannot be blinded to treatment assignment given the nature of the intervention, efforts are made to minimise bias through use of objective outcome measures and blinded data analysis. CLINICAL CONTEXT This study is embedded within routine clinical care for fibromyalgia patients at Oxford University Hospitals NHS Trust, Oxford Health NHS Foundation Trust, and Connect Health. Participants typically face approximately 6 months waiting time from diagnosis to receiving standard NHS pain rehabilitation treatment. During this waiting period, they can choose to participate in this research study. The study design thus takes advantage of an existing clinical gap to evaluate a potentially beneficial intervention while minimising delays to standard care. SCIENTIFIC INNOVATION This study represents several important innovations in chronic pain and sleep research: 1. Comprehensive mechanistic evaluation: Unlike previous studies, this trial combines physiological sleep measures (EEG), objective pain assessment (QST), and multiple modalities of movement measurement (VR, motion capture, actigraphy) to understand how sleep interventions affect pain and physical function. 2. Novel behavioural assessment tools: The VR paradigm and motion capture system provide ecologically valid, precise, and quantifiable measures of kinesiophobia and movement that have not been previously used in CBT-I trials for chronic pain. 3. Digital intervention consistency: Use of dCBT-I eliminates therapist variability and allows detailed monitoring of treatment engagement and adherence, facilitating better understanding of dose-response relationships. 4. Translational approach: The integration of basic science concepts (maladaptive learning) with clinical intervention testing provides potential insights into mechanisms underlying chronic pain and treatment response. 5. Patient-centred design: The VR task and motion capture system were co-designed with fibromyalgia patients, ensuring clinical relevance and acceptability. POTENTIAL IMPACT If dCBT-I proves effective in improving quality of life and other outcomes in fibromyalgia, it could provide a readily accessible, scalable intervention that could be rapidly implemented in clinical practice, as Sleepio is already NICE-accredited. Understanding the mechanisms through which sleep affects pain and movement could inform development of enhanced interventions and identification of patients most likely to benefit from sleep-focused treatments. The novel assessment tools developed and validated in this study could be applied to other chronic pain populations and intervention studies.