Effectiveness of a just-in-time adaptive intervention for improving pain catastrophising in patients with chronic back pain: protocol for a pragmatic randomised controlled trial Authors: Enya Sühs, Lotte Auguste Atkins, Lena-Sophie Koeppen, Paul Schmigalle, Charlott Scholz, Gregor Liegl, Rüdiger Pryss, Robin Kraft, Carsten Vogel, Abdul Rahman Idrees, Andreas Kopf, Anett Reisshauer, Peter Vajkoczy, Fabian Prasser, Sandra Nolte, Matthias Rose, Alexander Obbarius Published: 2026 ABSTRACT AbstractIntroductionChronic back pain (CBP), defined as back pain persisting for more than 3 months, is highly prevalent and associated with substantial socioeconomic burden and impaired quality of life. Pain catastrophising is an important cognitive-affective factor in the maintenance of CBP and may fluctuate in daily life. Conventional treatment approaches often do not address these momentary dynamics. Just-in-time adaptive interventions (JITAIs) deliver tailored support at opportune moments based on time-varying information and may therefore complement existing treatments for CBP. In this study, ecological momentary assessment (EMA) is used to capture momentary pain-related states and guide the delivery of adaptive microinterventions. To our knowledge, no JITAI specifically targeting pain catastrophising in CBP has yet been evaluated in a randomised controlled trial.Methods and analysisThis pragmatic, two-arm randomised controlled trial will compare a JITAI with enhanced treatment as usual (TAU+). A total of 100 participants will be recruited to obtain an evaluable sample of 86 participants. Eligible participants are adults with CBP persisting for more than 3 months and an average pain intensity over the previous week of at least 3 on a 0–10 Numerical Rating Scale. Participants in the intervention group will receive app-based, adaptive microinterventions, such as relaxation exercises and physical activities triggered by EMA responses on pain catastrophising and pain intensity. Participants in the TAU+ group complete the same EMA assessments but will not receive adaptive micro-interventions. The primary outcome is reduction in pain catastrophising, assessed with the Pain Catastrophizing Scale. Secondary outcomes include pain intensity, pain interference, pain behaviour, pain quality, physical functioning, anxiety, depression, fatigue, sleep disturbance, social participation, fear-avoidance beliefs, pain-related self-efficacy and treatment satisfaction. Primary and secondary outcomes will be analysed using between-group comparisons according to the intention-to-treat principle. Exploratory analyses will examine proximal, short-term effects of microinterventions using dynamic structural equation modelling. Assessments will be conducted at baseline before randomisation (T0), mid-treatment 2 weeks after intervention start (T1), end of treatment 4 weeks after intervention start (T2) and at 3 months (T3) and 6 months (T4) after end of treatment.Ethics and disseminationThis study has been approved by the Ethics Committee of the Charité Universitätsmedizin Berlin (approval number: EA4/207/21, 18 October 2021). Results will be submitted for publication in peer-reviewed journals and presented at conferences.Trial registration numberDRKS00028494.Keywords: Pain management, Back pain, Chronic Pain, Mobile Applications STRENGTHS AND LIMITATIONS OF THIS STUDY Ecological momentary assessment is used to repeatedly capture real-time pain-related experiences in participants’ natural environments. Ecological momentary assessment is used to repeatedly capture real-time pain-related experiences in participants’ natural environments. The intervention group includes micro-randomisation to explore short-term effects of different microinterventions. The intervention group includes micro-randomisation to explore short-term effects of different microinterventions. The pragmatic randomised controlled design compares the just-in-time adaptive intervention with enhanced treatment as usual under routine outpatient care conditions. The pragmatic randomised controlled design compares the just-in-time adaptive intervention with enhanced treatment as usual under routine outpatient care conditions. Participants and parts of the study team cannot be blinded because of the app-based intervention. Participants and parts of the study team cannot be blinded because of the app-based intervention. Generalisability may be limited by the requirement for smartphone access and the exclusion of patients with severe psychiatric or physical comorbidities. Generalisability may be limited by the requirement for smartphone access and the exclusion of patients with severe psychiatric or physical comorbidities. Introduction Chronic back pain (CBP), defined as back pain persisting for longer than 3 months, is highly prevalent and represents a major public health burden.1 A large population-based study reported an overall CBP prevalence of 38.3%.2 Back pain is a leading contributor to disability worldwide and causes substantial work-related burden, including sickness absence and long-term work disability.3 4 In addition to its socioeconomic consequences, CBP is associated with impaired quality of life, including limitations in physical, psychological and social functioning.5 6 Given its high prevalence, economic impact and detrimental effects, prevention and treatment strategies are needed that address modifiable factors contributing to the maintenance of CBP. Pain intensity, pain-related thoughts, affective states and behaviour often fluctuate over time and across everyday contexts. Conventional assessment and treatment approaches may insufficiently capture and address these within-person dynamics. Ecological momentary assessment (EMA) enables repeated real-time assessment of symptoms, behaviours, emotions and contextual factors in participants’ natural environments.7 Compared with retrospective questionnaires, EMA can capture momentary fluctuations and reduce recall bias.7 Just-in-time adaptive interventions (JITAIs) are personalised intervention strategies that deliver support at opportune moments based on time-varying information about an individual’s state, context or needs.7–10 JITAIs are often delivered via mobile or digital devices and can use different sources of time-varying information, including EMA, sensor data or usage patterns. In the present study, EMA is used to assess momentary pain-related states and to guide the delivery of adaptive microinterventions (MIs).9 This approach may allow interventions to be provided when patients are most likely to benefit from them. Pain catastrophising, a maladaptive cognitive and emotional response to pain, plays a critical role in the persistence and exacerbation of chronic pain.11 Characterised by feelings of helplessness, rumination and magnification of pain-related concerns, pain catastrophising has been identified as a key target for intervention.12 13 By assessing pain catastrophising in real time using EMA, an EMA-guided JITAI can deliver targeted strategies such as relaxation exercises, cognitive reframing or mindfulness practices when pain-related distress is elevated, thereby potentially interrupting the cycle of pain catastrophising and worsening pain.7 Because JITAIs are delivered in or near real time, often through mobile or digital devices, they can respond to changes in a patient’s context, behaviour or needs. This allows greater flexibility and responsiveness compared with traditional interventions delivered at fixed times. Several JITAIs have already been developed for various health-related purposes, including smoking cessation, physical activity promotion in cardiovascular disease and diabetes management, with meta-analytic evidence suggesting beneficial effects compared with non-JITAI interventions.14 To date, however, evidence for JITAIs in CBP remains limited. Related digital and EMA-based approaches have been applied in pain-related contexts, including opioid use disorder and CBP,15 but to our knowledge no randomised controlled trial (RCT) has evaluated an EMA-guided JITAI specifically targeting pain catastrophising in patients with CBP. A large proportion of patients with chronic pain benefit inadequately from existing treatments.16 Therefore, strategies are needed that provide support tailored to individual circumstances and momentary needs. Dynamic interventions delivered at times when they are most likely to be effective represent one approach toward more precise pain medicine.17 In addition to their potential clinical utility, EMA-guided JITAIs can generate high-resolution data that may improve understanding of the processes maintaining CBP. The study described in this protocol aims to evaluate the effectiveness of a novel EMA-guided JITAI for patients with CBP. The intervention combines repeated EMA with evidence-based MIs targeting pain catastrophising, pain intensity and related cognitive, emotional and behavioural factors. To achieve these aims, the following research questions will be examined: Does the JITAI reduce pain catastrophising compared with enhanced treatment as usual (TAU+)? Does the JITAI reduce pain catastrophising compared with enhanced treatment as usual (TAU+)? Does the JITAI improve secondary outcomes, including pain-related factors, emotional well-being, physical functioning, pain-related self-efficacy and fear-avoidance beliefs, compared with TAU+? Does the JITAI improve secondary outcomes, including pain-related factors, emotional well-being, physical functioning, pain-related self-efficacy and fear-avoidance beliefs, compared with TAU+? Do proximal changes in momentary outcomes, such as pain intensity and pain catastrophising, show exploratory associations with distal outcomes? Do proximal changes in momentary outcomes, such as pain intensity and pain catastrophising, show exploratory associations with distal outcomes? Which MIs, such as relaxation exercises, cognitive reframing, mindfulness-based techniques or physical activities, are associated with short-term improvements in proximal outcomes? Which MIs, such as relaxation exercises, cognitive reframing, mindfulness-based techniques or physical activities, are associated with short-term improvements in proximal outcomes? To what extent are app engagement indicators, such as adherence to EMA prompts and MIs, descriptively associated with clinical outcomes? To what extent are app engagement indicators, such as adherence to EMA prompts and MIs, descriptively associated with clinical outcomes? Methods and analysis Design This is a pragmatic, two-arm RCT comparing a JITAI group with an enhanced TAU+ group. Pragmatic trials aim to evaluate how well an intervention works under usual care conditions, with broader inclusion criteria and flexible implementation to enhance real-world applicability.18 Pragmatic features of the present study include recruitment from outpatient clinics, delivery of the intervention in participants’ everyday lives and unrestricted access to usual care in both study groups. Patients will be enrolled at the clinic, where they will receive an invitation to attend a practice session and an introduction to the app. The baseline assessment will be conducted on-site using a mobile application. A detailed technical description of the mobile application framework used to develop the BackUp mobile application has been published previously.19 20 Assessments will be conducted at baseline before randomisation (T0), mid-treatment at 2 weeks after intervention start (T1), end of treatment at 4 weeks after intervention start (T2), and at 3-month (T3) and 6-month follow-up (T4) after end of treatment. Baseline assessments will be conducted on site, while all follow-up assessments will be completed independently by participants online via Research Electronic Data Capture (REDCap). The study started recruitment on 1 January 2026. Participant recruitment is planned to continue until December 2026. The final 6-month follow-up assessment is expected to be completed by July 2027. A flow chart of the study procedure is shown in figure 1. Flow chart of the study procedure. Flow chart illustrating participant recruitment, screening according to the predefined inclusion and exclusion criteria, randomisation to the JITAI intervention or treatment-as-usual plus (TAU+) group, and the schedule of baseline, postintervention and follow-up assessments. iOS, Apple’s mobile operating system; JITAI, Just-in-Time Adaptive Intervention; PHQ-9, Patient Health Questionnaire-9. Inclusion and exclusion criteria Patients eligible for participation must meet the following inclusion criteria: (a) age 18 years or older, (b) proficiency in German, including speaking, reading, and writing, (c) CBP, including lumbar and/or thoracic back pain, persisting for ≥3 months, (d) average pain intensity over the previous week of at least 3 on a 0–10 numerical rating scale and (e) ownership of a smartphone (Android or iPhone). Participants will be excluded if they (a) are unable to comprehend the study information, (b) have moderate to severe depression (Patient Health Questionnaire-9 (PHQ-9) score ≥15), (c) report suicidality (PHQ-9 Item 9≥1), or have a history of severe suicidality requiring treatment, (d) present with radicular symptoms that have not been medically clarified (pain in the legs/feet, muscle weakness in the legs/feet, sensory disturbances such as tingling/numbness in the legs/feet), (e) have had knee replacement surgery, (f) experience a dizziness disorder, (g) have a rheumatic disease, (h) have an indication for back surgery, (i) exhibit significantly limited mobility (eg, unable to perform exercises on a mat) or (j) suffer from severe mental or physical illness requiring treatment, such as psychotic disorder, eating disorder, post-traumatic stress disorder, cancer, exacerbated chronic obstructive pulmonary disease or acute coronary syndrome. Recruitment and procedure Participants are being recruited from outpatient clinics at Charité – Universitätsmedizin Berlin. Recruitment strategies include displaying posters in outpatient clinics and direct referrals by treating physicians. Eligible participants will be invited to provide written informed consent (see online supplemental file 1, Participant Consent Form) before proceeding to the baseline assessment (T0). After completing the baseline assessment, participants will be introduced to the smartphone application. Participants will then be randomly assigned to either the JITAI group or the enhanced TAU+ group. Both groups will be prompted to complete EMA surveys five times daily. Survey prompts will occur randomly within prespecified time windows throughout the day to reduce predictability and capture natural fluctuations in participants’ experiences7: 08:00–08:30, 09:00–12:00, 12:30–15:30, 16:00–19:00 and 19:30–20:00. These windows are used to reduce predictability while ensuring repeated assessment during waking hours. For each prompt, participants will receive one reminder after 30 min if the survey has not been completed. Only the JITAI group will receive personalised MI prompts based on app responses. Participants who miss more than 2 days of EMA surveys will be contacted by the study team via email or phone to encourage adherence. Follow-up assessments (T1–T4) will be conducted via REDCap with automated notifications sent to remind participants to complete these assessments.21 Randomisation and blinding Participants will be randomly assigned to either the JITAI group or the TAU+ group using permuted block randomisation with variable block sizes (2, 4 or 6) and an allocation ratio of 1:1.22 The randomisation list will be generated using the R package ‘simstudy’ and securely managed through REDCap to ensure allocation concealment.23 Due to the nature of the intervention, blinding of study participants is not feasible, as participants in the JITAI group will receive MI suggestions based on EMA survey responses, while the TAU+ group will not.24 At least one member of the study team will also remain unblinded to remind participants about app-based tasks and encourage adherence to the protocol. Since participants are unblinded, no unblinding procedures are necessary. Participants may be withdrawn from the study if they no longer meet the inclusion criteria, fail to adhere to the protocol or request withdrawal. Micro-randomisation Micro-randomisation involves the random allocation of intervention options at relevant decision points, defined as moments when a specific intervention component is likely to be effective.25 In this study, micro-randomisation is implemented within the JITAI group. When predefined decision rules based on EMA responses indicate that an MI should be delivered, the type of MI is randomly selected from the available intervention options, such as relaxation exercises, cognitive reframing or physical activities. By repeatedly randomising intervention components at the individual level, micro-randomised designs can support the estimation of proximal effects of specific intervention components while accounting for temporal and contextual variability.25 26 In the present study, the proximal response to an MI is assessed at the subsequent EMA prompt. Proximal effects are therefore defined as short-term changes in momentary outcomes, such as pain intensity or pain catastrophising, from one EMA prompt to the next following MI delivery. These micro-randomised data will be used in exploratory analyses to examine which MIs may be more suitable for specific patients or situations. The study is powered for the primary distal RCT outcome, not for confirmatory comparisons between individual MIs. Power calculation and sample size The sample size calculation was based on the primary distal RCT outcome, reduction in pain catastrophising as measured by the Pain Catastrophising Scale (PCS). Assuming a medium effect size of f=0.25,12 27 80% power, a two-sided significance level of α=0.05, and two treatment groups, an evaluable sample of 86 participants was estimated to be required using G*Power.28 Accordingly, approximately 43 evaluable participants are required per group. To account for an expected dropout rate of 15%, 100 participants will be recruited and randomised. Intervention development The intervention was designed based on previous recommendations for JITAIs9 and developed using a patient-centred approach. The development process included a targeted literature review of self-management strategies for CBP, expert input from physical medicine, pain treatment, psychosomatic medicine, and psychotherapy, and structured interviews with 15 patients with CBP and 15 pain experts. These steps informed the selection and prioritisation of MIs from the domains of relaxation, psychotherapy and physiotherapy, as well as design elements such as intervention frequency, tailoring variables, decision thresholds and motivational features. Findings were synthesised into a design document specifying decision points, tailoring variables, including pain intensity and pain catastrophising and decision rules for triggering MIs. A detailed description of the development process is provided elsewhere.29 Patient and public involvement Patients with CBP were involved in the development of the intervention. Structured interviews with 15 patients were conducted to assess the feasibility and anticipated usefulness of candidate MIs and to refine design elements such as intervention frequency, decision thresholds and motivational features. Patient input informed the selection and delivery of the MIs. Patients were not involved in the design of the randomised trial or the statistical analysis plan. Intervention content The selected MIs are brief, practical strategies that aim to address the psychological, physical and behavioural dimensions of CBP. Using the eSano platform,30–32 they draw on evidence-based techniques from relaxation therapy and cognitive behavioural approaches including acceptance and commitment therapy and physical activity interventions. Each intervention is designed to fit seamlessly into daily life and requires minimal time and effort to implement (eg, a maximum of 10–15 min per MI). The focus is on promoting immediate relief from catastrophising and other pain-related distress (proximal outcomes) while fostering long-term improvements in pain management and quality of life (distal outcomes). Interventions are designed to help participants manage pain catastrophising, reduce emotional distress and improve mobility. For example, relaxation-based interventions provide guided exercises to promote calmness and reduce muscle tension, while cognitive techniques focus on reframing negative thought patterns related to pain. Physical activity interventions are tailored to participants’ abilities and encourage gentle movement to counteract stiffness and improve physical function. To ensure usability, all interventions are interactive and include step-by-step instructions, audio guidance or visual demonstrations. TAU+ content Participants in the TAU+ group complete the same EMA surveys as the JITAI group but do not receive personalised MI prompts and are not introduced to the MIs. The TAU+ group retains unrestricted access to standard care options. Participants are monitored for adherence to EMA prompts throughout the study to ensure comparable data collection between groups. Outcomes This study includes distal primary, distal secondary and proximal outcomes. Baseline covariates are collected to describe the sample and explore potential subgroup differences in treatment effects. Distal outcomes refer to clinical outcomes assessed using recall-based questionnaires, whereas proximal outcomes refer to short-term, momentary pain-related experiences assessed repeatedly in real time using EMA. The primary distal outcome is reduction in pain catastrophising, measured by the PCS. Secondary distal outcomes include pain intensity, pain interference, pain behaviour, pain quality, physical functioning, anxiety, depression, fatigue, sleep disturbance, participation in social roles and activities, fear-avoidance beliefs, pain-related self-efficacy and treatment satisfaction. Proximal outcomes include momentary pain catastrophising, pain intensity, stress, depressive symptoms, fatigue and positive affect. Proximal outcomes following a micro-randomised intervention are assessed at the subsequent EMA prompt. The theoretical rationale of the intervention is that repeated short-term improvements in proximal outcomes, particularly reductions in momentary pain catastrophising and pain intensity, may accumulate over time and contribute to distal improvements in CBP-related functioning and well-being. Distal primary outcome The primary outcome is the reduction in pain catastrophising, which is assessed using the PCS.13 The PCS evaluates three subdomains: rumination (persistent negative thoughts about pain), magnification (exaggerated perceptions of pain threat) and helplessness (a sense of lack of control over pain). Responses are rated on a 5-point Likert scale ranging from ‘not at all’ to ‘all the time’. The PCS has been validated and demonstrates excellent internal consistency (Cronbach’s α=0.92) and acceptable test–retest reliability (r=0.88). The primary outcome captures the intervention’s impact on a cognitive factor that perpetuates chronic pain. Distal secondary outcomes Patient-Reported Outcomes Measurement Information System pain measures The Patient-Reported Outcomes Measurement Information System (PROMIS) is a standardised system for measuring health-related quality of life across various domains.33 In this study, we will use fixed-length short forms that allow efficient and precise assessment while minimising participant burden (https://www.assessmentcenter.net/). PROMIS measures in this study use Likert-style responses, mostly 5-point. The pain intensity item is rated on an 11-point numeric rating scale. All PROMIS scores are reported on a T-score metric, where 50 represents the mean of the US general population with an SD of 10. Pain-related outcomes are assessed using the PROMIS pain interference Short-Form 8a v1.1, which evaluates the degree to which pain limits engagement in daily activities, work and social interactions.34 35 The PROMIS pain intensity item (Global07) is used as a single-item measure of pain severity.36 The PROMIS pain behaviour Short-Form 7a v1.0 measures observable pain-related behaviours such as grimacing, guarding or movement avoidance.37 Pain quality is assessed using the PROMIS neuropathic pain quality short-form 5a v2.0 and the PROMIS nociceptive pain quality short-form 5a v2.0, capturing sensory characteristics of pain, including descriptors such as tingling, burning, sharpness or aching sensations.38 Other PROMIS measures In addition to pain-related outcomes, several PROMIS domains are included to assess psychological distress, physical functioning, fatigue, sleep quality and social participation.39 The PROMIS physical function short-form 8b v2.0 evaluates participants’ ability to perform daily tasks requiring mobility and strength. Psychological distress is assessed using the PROMIS anxiety short-form 4a v1.0 and the PROMIS depression short-form 4a v1.0, measuring symptoms such as excessive worry, nervousness, hopelessness and low mood. Fatigue is measured using the PROMIS fatigue short-form 4a v1.0, which evaluates perceived energy levels and the extent to which fatigue limits engagement in daily activities. Sleep quality is assessed using the PROMIS sleep disturbance short-form 4a v1.0, which captures perceived sleep quality, difficulties falling or staying asleep and overall restfulness. The PROMIS ability to participate in social roles and activities short-form 8a v2.0 is used to measure restrictions in social and occupational engagement. Fear avoidance beliefs The Fear Avoidance Beliefs Questionnaire (FABQ) assesses the extent to which individuals with chronic back pain avoid physical activity and work-related tasks due to fear of pain exacerbation40–42. It consists of 16 items, divided into two subscales: physical activity avoidance (items 2–5) and work-related avoidance (items 6, 7, 9-12 and 15). Responses are recorded on a 7-point Likert scale ranging from 0 ("Not true at all") to 6 ("Completely true"). The FABQ has demonstrated good internal consistency (Cronbach’s α = 0.82) but questionable test-retest reliability (r = 0.64), indicating potential variability in responses over time. Pain-related self-efficacy Pain-related self-efficacy reflects an individual’s confidence in their ability to engage in daily activities and achieve personal goals despite experiencing pain. Pain-related self-efficacy is assessed using the German adaptation of the PSEQ, the Fragebogen zur Erfassung der schmerzspezifischen Selbstwirksamkeit (FESS).43 The FESS comprises 10 items rated on a six-point scale, yielding a total score ranging from 10 to 60. Higher scores indicate greater pain-related self-efficacy. The PSEQ has been validated across various chronic pain populations, demonstrating acceptable to excellent internal consistency (Cronbach’s α=0.79–0.95) and good test–retest reliability (intraclass correlation coefficient (ICC)=0.75–0.93), supporting its stability over time.44 Personality functioning The Operationalised Psychodynamic Diagnosis (OPD)-Structure Questionnaire Short Form (OPD-SQS) is a 12-item self-report measure assessing impairments in personality structure based on the OPD model. It covers three core dimensions of structural integration: self-perception, interpersonal contact and relationship models. Items are rated on a five-point scale from 0 (“fully disagree”) to 4 (“fully agree”), yielding a total score ranging from 0 to 48, with higher scores indicating greater impairment in personality functioning. The OPD-SQS has demonstrated good internal consistency (Cronbach’s α=0.82–0.87) and a stable three-factor structure across diverse samples.45 A summary of recall assessments is shown in table 1. Table 1Summary of recall assessmentsInstrumentTime of measurementScreeningT0T1T2T3T4Screening instruments PHQ-9Patient Health Questionnaire-9 (Depression)X Numeric Rating ScalePROMIS pain intensity (Global07)XPrimary outcome PCSPain Catastrophising ScaleXXXXXSecondary outcomes PROMISPain intensity item (Global07)XXXXX Pain interference Short Form 8a v1.1XXXXX Pain behaviour Short Form 7a v1.0XXXXX Neuropathic pain quality Short Form 5a v2.0XXXXX Nociceptive pain quality Short Form 5a v2.0XXXXX Physical function Short Form 8b v2.0XXXXX Anxiety Short Form 4a v1.0XXXXX Depression Short Form 4a v1.0XXXXX Fatigue Short Form 4a v1.0XXXXX Sleep disturbance Short Form 4a v1.0XXXXX Ability to participate in social roles and activities Short Form 8a v2.0XXXXX FABQFear Avoidance Beliefs QuestionnaireXXXXX PSEQPain Self-Efficacy QuestionnaireXXXXX OPD-SQSOPD Structure Questionnaire Short Form (Personality functioning)XXXOther assessments RPSRegional Pain Scale (Pain location)X CTQChildhood Trauma QuestionnaireX Satisfaction with treatmentXXX AEAdverse EventsXXXXPROMIS, Patient-Reported Outcomes Measurement Information System. Summary of recall assessments PROMIS, Patient-Reported Outcomes Measurement Information System. Proximal outcomes Participants receive five daily EMA prompts at random times within the predefined windows described above. Proximal outcomes following a micro-randomised intervention are assessed at the subsequent EMA prompt. Analyses of these proximal effects are exploratory, whereas the study was powered for the primary RCT outcome. These data allow for the evaluation of short-term intervention effects, examining how MIs influence immediate pain perceptions, affective states and cognitive responses. An overview of the momentary items is provided in table 2. Table 2Summary of momentary assessmentsTimeConstructValenceContentMorning questionsSleep qualityPositiveLast night, I slept well.Sleep restedPositiveAs I woke up, I felt well rested.Outlook: positivePositiveI look forward to the rest of the day with hope.NegativeI don’t think I can take it anymore.Daily questionsPain catastrophising Short Form 1NegativeThe condition is terrible and I don't think it will ever get better.NegativeI am afraid that the pain will get worse.NegativeI think about how much it hurts.Pain catastrophising Short Form 2NegativeI’m worried about whether the pain will ever stop.NegativeI have the feeling I can't take it anymore.NegativeI think of other situations in which I was in pain.NegativeI think about how much I want the pain to end.Pain catastrophising Short Form 3NegativeThe condition is terrible and threatens to overwhelm me.NegativeI desperately want the pain to go away.NegativeI can’t stop thinking about the pain.NegativeI’m worried that the pain is a sign of something bad.Pain intensityNegativeHow would you rate your back pain right now?StressNegativeI’m feeling stressed right now.DepressionNegativeI feel depressed right now.DepressionNegativeI feel helpless right now.FatigueNegativeI’m tired right now.FatigueNegativeI’m exhausted right now.CheerfulPositiveI feel happy/cheerful right now.Evaluation of last microintervention*PositiveHow much did the last microintervention help you?Other interventionsNeutralHave you done anything about your pain since the last prompt?Activity categoryNeutralCurrent activityActivity ratingPositiveI enjoy what I’m doing right now.SocialNeutralI am currently spending time with…Social ratingPositiveI enjoy my company.LocationNeutralI am currently located…Evening questionOther painNeutralDid you have pain today that did not affect your back?*Only intervention group. Summary of momentary assessments Only intervention group. Pain intensity is assessed using a 0 to 10 NRS, in which participants rate their current back pain severity. Most momentary symptom and affect items are rated on five-point response scales, whereas contextual items use categorical response options. Momentary pain catastrophising is measured using three 4-item short forms adapted from the PCS—adaptive momentary PCS (EMA-PCS). These short forms were developed using item response theory methodology to ensure optimal psychometric properties and reduce response burden. A detailed description of their development and validation is reported elsewhere.46 Each short form includes at least one item from each PCS domain—helplessness, magnification and rumination—capturing key cognitive-affective aspects of pain catastrophising. To minimise item exposure bias,47 the short forms rotate across EMA prompts, ensuring that participants do not repeatedly answer the same questions. Momentary stress, depressive symptoms and fatigue are assessed using items adapted from PROMIS measures and previous EMA research.48 EMA surveys also assess current activity, social context and location to account for situational influences on pain and psychological states. Participants in the intervention group additionally evaluate the perceived helpfulness of the most recent MI using a single-item rating scale from 0 (‘not at all helpful’) to 4 (‘very helpful’). Participants also report whether they engaged in other pain management strategies since the previous prompt, such as medication use, relaxation exercises or physical activity. Intervention-related variables Adherence and app engagement will be assessed using objective app usage data, including completed MIs, response rates to EMA prompts and dropout rates. These data will be collected automatically via the intervention platform in compliance with the European General Data Protection Regulation and institutional data-sharing agreements. Engagement indicators will be reported descriptively and explored in hypothesis-generating analyses of associations with clinical outcomes, without causal interpretation. Potential negative effects of the intervention will be assessed using structured open-ended questions. Participants will be asked to describe any negative events, including type, onset, frequency and duration. Additional Likert-scale questions will assess the perceived negative impact of these events at the time of occurrence and retrospectively at follow-up. Participants’ app usability and subjective experience will be assessed using the Mobile App Rating Scale (uMARS).49 Because no German version of the uMARS is available, the MARS-G50 will be used while retaining the original uMARS items where applicable. Baseline covariates Baseline covariates are collected to describe the sample and explore potential subgroup differences in treatment effects. Demographic variables include age, gender, educational level, employment status and living situation. Clinical variables include pain duration, previous surgeries, comorbid conditions, substance use and current medication. Behavioural and psychosocial variables include physical activity, personality functioning and childhood trauma history assessed with the Childhood Trauma Questionnaire.51 Screening instruments Patient Health Questionnaire-9 The PHQ-952 is used to screen for depressive symptoms and suicidality. It consists of nine items rated from 0 (‘not at all’) to 3 (‘nearly every day’), yielding a total score from 0 to 27. In this study, participants with a PHQ-9 score≥15 or any indication of suicidality on item 9 are excluded for safety reasons. Pain intensity Average pain intensity over the previous week is assessed using an 11-point Numerical Rating Scale53 ranging from 0 (‘no pain’) to 10 (‘worst imaginable pain’). Participants are eligible if they report an average pain intensity of at least 3. Analyses All primary and secondary outcome analyses will follow the intention-to-treat principle. We will not conduct a confirmatory per-protocol analysis because adherence to EMA prompts and MIs is assessed after randomisation and may be associated with time-varying prognostic factors, which could bias naïve per-protocol estimates. Adherence and app engagement will be reported descriptively and explored in secondary, hypothesis-generating analyses without causal interpretation. Handling of missing data Patterns of missing data will be described, and potential predictors of missingness will be explored. Mixed-effects models will be used for longitudinal outcome analyses, as these models can include participants with incomplete repeated outcome assessments under a missing-at-random assumption. If appropriate, multiple imputation by chained equations will be used in sensitivity analyses to assess the robustness of findings under different assumptions about missingness.54 Distal outcome analyses The primary distal analysis will evaluate whether pain catastrophising changes differently over time in the JITAI group compared with the TAU+ group. PCS scores will be analysed using a mixed-effects model with fixed effects for group, time, and the group-by-time interaction, and a random intercept for participant to account for repeated measurements within individuals.55 Time will be modelled as a categorical variable according to the prespecified assessment points. The primary treatment effect will be evaluated using the group-by-time interaction for PCS. Prespecified model-based contrasts will be used to estimate between-group differences in change from baseline to end of treatment and to follow-up assessments. Secondary distal outcomes will be analysed using analogous mixed-effects models. These analyses will examine whether potential intervention effects extend to pain intensity, pain interference, pain behaviour, pain quality, physical functioning, anxiety, depression, fatigue, sleep disturbance, participation in social roles and activities, fear-avoidance beliefs and pain-related self-efficacy. Proximal outcome analyses Exploratory proximal analyses will be conducted within the JITAI group to examine which MIs are associated with short-term improvements in momentary outcomes and whether these effects differ between patients or across situations. Proximal outcomes following MI delivery will be assessed at the subsequent EMA prompt. Dynamic structural equation modelling and related multilevel time-series models will be used to account for the nested structure of EMA data, with repeated observations nested within participants.55 56 Models will include MI type as a time-varying predictor and will adjust for relevant prior-state variables, including the previous level of the outcome. Time-varying covariates, such as current pain intensity, pain catastrophising, affective state, fatigue, activity, social context, location and time of day, will be examined as potential moderators of proximal MI effects.26 Person-level characteristics, such as baseline pain duration, psychological distress, fear-avoidance beliefs, and pain-related self-efficacy, will be explored as between-person moderators. These analyses are exploratory and intended to inform future optimisation of JITAI decision rules. The study was powered for the primary distal RCT outcome and not for confirmatory comparisons between individual MIs, as sample size requirements for micro-randomised proximal effects depend on separate assumptions, including the number of decision points, randomisation probabilities, availability, expected proximal effect size and within-person correlation.57 Exploratory moderator and mechanism analyses Additional exploratory analyses will examine whether baseline characteristics are associated with differential distal treatment effects. Hypothesis-generating analyses will also examine whether proximal changes in momentary pain catastrophising, pain intensity or stress are associated with distal clinical outcomes. Discussion This protocol describes a pragmatic RCT evaluating an EMA-guided JITAI for patients with CBP. The intervention combines repeated momentary assessment with adaptive delivery of brief MIs targeting pain catastrophising, pain intensity and related cognitive, emotional and behavioural factors in daily life. The primary aim is to evaluate whether the JITAI reduces pain catastrophising compared with TAU+. In addition, the intensive longitudinal EMA data will allow exploratory analyses of short-term changes in pain-related states and of whether specific MIs may be more suitable for particular patients or situations. Several methodological strengths should be noted. The pragmatic design, recruitment from outpatient clinics and unrestricted access to usual care are intended to increase the clinical relevance of the findings.58 The use of EMA enables repeated assessment of pain-related experiences in participants’ natural environments and reduces reliance on retrospective reports.7 The micro-randomised component within the JITAI group provides an opportunity to explore proximal effects of different MIs and to generate hypotheses for future optimisation of JITAI decision rules.9 25 The intervention was developed using a patient-centred process that included patient and expert input, supporting the clinical plausibility and usability of the intervention components.29 The study also has limitations. Participants and parts of the study team cannot be blinded because of the app-based intervention and adaptive delivery of MIs, which is a common challenge in trials of non-pharmacological and behavioural interventions.24 Generalisability may be limited by the requirement for smartphone access, sufficient German language proficiency and the exclusion of patients with severe psychiatric or physical comorbidities. These criteria may reduce heterogeneity and support safe study conduct, but they may also limit inclusion of patients who experience barriers to digital health participation.59 60 In particular, participants with moderate to severe depression or suicidality are excluded for safety reasons, because the intervention includes self-guided components without continuous therapeutic supervision and is not specifically designed as a suicide-prevention or crisis-support intervention. Although digital interventions targeting suicidal ideation may be effective when specifically developed and evaluated for this purpose, their use requires appropriate safety procedures and further evidence regarding suicidal behaviour and relevant moderators.61 62 Consequently, findings may not generalise to patients with CBP who require intensive psychiatric or psychotherapeutic care. Future studies should examine whether JITAIs can be safely adapted for populations with higher psychological distress under closer clinical supervision. Another limitation concerns the exploratory nature of the proximal analyses. Micro-randomised designs can support the estimation of proximal effects of specific intervention components and help investigate whether these effects vary across time, contexts or patient characteristics.25 26 However, sample size requirements for micro-randomised proximal effects depend on assumptions that differ from those used for the primary distal RCT outcome, including the number of decision points, availability, randomisation probabilities, expected proximal effect size and within-person correlation.57 The present study was powered for the primary distal RCT outcome and not for confirmatory comparisons between individual MIs. Analyses examining which MIs are associated with short-term improvements in specific patients or situations will therefore be interpreted as hypothesis-generating. In addition, adherence to EMA prompts and MIs may vary between participants and over time, which is a common challenge in digital health interventions.63 64 App engagement will therefore be reported descriptively and explored in relation to clinical outcomes without causal interpretation. If effective, EMA-guided JITAIs may complement existing treatments for CBP by providing brief, adaptive support in daily life. Beyond evaluating clinical outcomes, this study may contribute to a better understanding of within-person dynamics in pain catastrophising and pain intensity and may inform the development of more precise intervention strategies for CBP. Ethics and dissemination This study has been approved by the Ethics Committee of the Charité – Universitätsmedizin Berlin (approval number: EA4/207/21). All participants will provide written informed consent before enrolment. The study will be conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Participants’ privacy and confidentiality will be strictly maintained throughout the study. Data will be pseudonymised and stored securely, with access limited to authorised research personnel. Data processing will follow the EU General Data Protection Regulation and the Berlin State Data Protection Act. Participants may withdraw from the study at any time without consequence. The results of this study will be disseminated through peer-reviewed publications and presentations at relevant scientific conferences. A summary of the findings will also be made available to study participants on request. Trial status Recruitment for this study started on 1 January 2026 and is currently ongoing. The JITAI app, participant training materials and REDCap infrastructure have been implemented for study conduct. The study is registered with the German Clinical Trials Register (DRKS00028494; registered on 10 March 2022). Supplementary material Acknowledgements AO is participant in the BIH Charité Digital Clinician Scientist Program funded by the Charité – Universitätsmedizin Berlin and the Berlin Institute of Health at Charité (BIH). The funders had no role in the design of the study, collection, management, analysis and interpretation of data, writing of the manuscript or the decision to submit the manuscript for publication. Footnotes Funding: This work was supported by Charité – Universitätsmedizin Berlin and the Berlin Institute of Health at Charité. Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2025-110437). Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Patient and public involvement: Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details. References Supplementary Materials