Chronic pain is a major contributor to worldwide disability, affecting approximately 20% of the global population [1-3]. For many patients, ongoing or recurrent pain severely impacts their physical, social, and mental health, as it interrupts ongoing activities and thereby continuously interferes with daily life functioning. Over time, this impacts patients’ sense of self and quality of life [4].

In many cases, there is no monodisciplinary treatment available that can cure the persisting pain. Instead, the multifaceted nature of chronic pain, including biomedical as well as psychological and social factors, often requires a comprehensive treatment approach focusing on improving daily life functioning rather than reducing pain [5-7]. To realize this, interdisciplinary multimodal pain therapy (IMPT) programs have evolved that aim to support patients in learning to live a meaningful life irrespective of pain. These programs share a biopsychosocial orientation toward chronic pain and often include both neuroscientific models of pain physiology as well as (cognitive) behavioral treatment principles [6,8].

Although the effectiveness of IMPT programs has been well established [9-12], maintaining the positive effect of the treatment on patients’ daily lives over time remains a major challenge [13,14]. The problem of relapse is not unique to the domain of pain treatment but has been observed across all health behavior domains (eg, [15,16]). In response, many treatment programs have added relapse prevention strategies that aim to preserve treatment gains over time (eg, [17,18]). In the context of chronic pain treatment, examples of such strategies include self-practice exercises [19,20], booster sessions [21,22], or encouragement of patients to take notes during treatment [23,24]. However, the integration of these particular strategies within the treatment program as well as an underlying theoretical rationale regarding how it may prevent relapse are often not described in clinical studies. Moreover, the effectiveness of these behavior regulation strategies remains unknown because they are usually evaluated as a part of the full program or with a limited follow-up period.

The emergence of mobile health (mHealth) provides new opportunities to support behavior regulation to maintain or enhance the long-term treatment effect of IMPT programs. Despite substantial variations concerning study quality, interventions, and outcomes, mHealth apps are generally regarded as a promising strategy to facilitate adherence to treatment principles or to increase self-management skills (eg, [25,26]). A specific advantage is that an app can include multiple interacting behavior regulation strategies within its digital environment (eg, automatically linking personalized goal setting regarding physical activity to accelerometer output). Moreover, mHealth strategies can integrate other smartphone functionalities such as digital calendars, instant messaging services, or a camera, thereby offering personalized behavior regulation strategies that support the transfer of treatment insights into each patient’s personal environment.

Despite the potential and current popularity of mHealth apps, the effectiveness on health-related outcome measures varies greatly [27]. Factors such as engagement—defined as the extent of app use as well as the corresponding subjective experience [28]—and usability—defined as the relative ease with which users can use an app to achieve a particular goal [29,30]—may account for this variability [27,30]. For example, patients that use an app to change their health behaviors may use the app in a different way than intended or stop using the app after several days, which prevents facilitating the intended behavior change [28]. Therefore, evaluations concerning the effectiveness and clinical importance of mHealth apps on health outcomes should take evaluations on user engagement and perceived usability into account [30,31].

In 2015, we initiated the SOLACE research project to develop strategies to prevent relapse after IMPT programs. Because there was little research available on this topic [13,14], we started with an 18-month co-design project, in which patients, health care providers, researchers, and designers shared their expertise and collaborated to develop ideas, concepts, and strategies to prevent relapse after successful treatment. This resulted in a prototype paper workbook that contained the two most promising strategies: a valued-based goal-setting procedure, and a method for storing and facilitating retrieval of meaningful treatment experiences. Subsequently, we performed a feasibility study in which the prototype workbook was tested at two different IMPT programs for 6 months. Overall, patients and health care providers were willing and able to use the workbook and regarded the strategies to be in line with the IMPT treatment principles (see personal communication, first author SE; manuscript under review). The evaluations also yielded specific suggestions for further improvements, including a preference for a mobile app instead of a paper workbook, along with more interaction between both strategies and a modified goal-setting procedure. In the ensuing research project (ie, the AGRIPPA project), we used the insights of the feasibility study to improve both strategies and transferred the content of the workbook prototype to an mHealth app. Similar to the initial co-design project, patients, health care providers, designers, and researchers collaborated to optimize usability and intervention components of the app. For example, we explored direct user experience through think-aloud sessions with digital mockups, and we organized cocreation sessions to prioritize app features and to prepare a list of requirements.

The present trial has three main objectives. Our first objective is to evaluate the effectiveness of the AGRIPPA intervention (ie, enhanced treatment condition) on pain disability for patients with chronic musculoskeletal pain who participate in an IMPT program compared to a usual care control group with a follow-up period of 12 months. Our second objective is to determine the cost-effectiveness of the AGRIPPA intervention relative to usual care. Our third objective is to explore the level of engagement and perceived usability of patients that use the app. We have formulated three hypotheses: (1) the maintenance of improvement in pain-related disability over time after IMPT will be more favorable for the enhanced treatment condition compared to usual care; (2) the effect of the app on pain disability will translate to less health care utilization and less societal costs (eg, less absenteeism), leading to a cost-effective intervention compared to treatment as usual; and (3) for participants in the enhanced treatment condition, the perceived usability, frequency of use, duration of use, and reported adherence to the AGRIPPA app will be positively associated with a favorable change over time of pain-related disability.

We will perform a randomized controlled multicenter superiority trial with two parallel groups in the Netherlands. Both groups will receive standard IMPT, but the experimental group will be provided with the AGRIPPA app that they can access both during and after the treatment program. The nature of the intervention does not allow for masking the condition for health care providers or patients. The allocation ratio will be 1:1.

The study activities have been reviewed and approved by the Medical Research Ethics Committee Utrecht (19/406/D). Protocol modifications that will result in significant changes of study objectives, design, or procedures will require approval by the AGRIPPA steering committee and the Medical Research Ethics Committee.

The trial will be coordinated by a senior researcher of the Lifestyle & Health Research Group of the University of Applied Sciences Utrecht. The AGRIPPA consortium that consists of all project partners (including the participating treatment centers) meets twice per year to discuss overall progress and topics such as dissemination. The AGRIPPA steering committee consists of researchers who meet four times per year to oversee the quality of the research and to decide on any substantial amendments to the initial project idea.

Six treatment facilities that provide IMPT programs participate in this study (2 hospital units and 4 rehabilitation clinics). All locations provide an interdisciplinary biopsychosocial-oriented treatment program to patients within the region, leading to a mixed rural and urban population throughout the Netherlands (ie, Arnhem, Eindhoven, Hoensbroek, Roermond, Maastricht, and Wijk aan Zee).

All patients who will start an IMPT program during the study inclusion period will be contacted by their treating physician in rehabilitation medicine and will receive a patient information letter. The patient will then have 1 to 2 weeks to consider participation and to ask any additional questions before signing the informed consent form. Instruction of the app will follow when a signed informed consent form has been provided to the research assistant. An independent research counselor will be available to all participating patients for any questions and general support during the study. Her email address will be provided via the patient information letter.

In participating treatment centers, treatment programs either consist of group sessions, individual sessions, or their combination. To prevent contamination, we will perform group randomization when treatment is predominantly provided in groups. When treatment is provided in individual sessions, we will randomize each patient individually using a simple randomization procedure. The randomization will be performed using random allocation cards based on computer-generated random numbers. The randomization procedure will be executed by a research assistant who is not providing treatment, and will be concealed from health care providers and patients using a set of sequentially numbered opaque and sealed envelopes. The group randomization procedure will be performed in a similar manner for each group that contains at least one participating patient.

All six treatment facilities routinely collect outcome data with an electronic survey to monitor their patients [33]. All demographics, baseline measurements, as well as the primary outcome will be obtained by this procedure. Any outcome measures that are not part of the routine assessment will be obtained through an additional electronic survey. Table 1 includes an overview of all outcome domains, measures, and planned analysis methods. The participant flow that is depicted in Figure 2 includes all time points of data collection in this study. A researcher who is not involved with the treatment program will monitor all incoming data and will promote study retention using email reminders after 7 days and telephone reminders at 15 and 21 days postmeasurement.

All study data will be obtained via three electronic sources. The questionnaires will either be collected through routine monitoring procedures within the treatment centers or via additional electronic surveys. System usage data will be provided by the app developer. All study data will be stored within the firewall of the University of Applied Sciences Utrecht (UAS) in a folder on a network drive that is protected by permission rights and will only be available to researchers that are assigned by the project team to analyze the data. Data will be automatically backed up (daily) by the UAS Utrecht. To protect the identity of individual participants, we will perform the following procedures for personally identifiable information. During the data collection phase, we will pseudonymize all incoming data. In the main dataset, we will replace identity data with a unique number (ie, identifier). Date of birth will be transformed to age in years and address information will not be included in the dataset, except for province and place of residence (rural/urban). To add additional measurements to the dataset and to delete data upon participant request, we will create a correspondence table that contains both the identifier and patient personal information. This table will be stored in a separate folder that only the principal investigator (JP) and the head of the Lifestyle & Health Research Group (HW) can access. Any unforeseen data collection issues that may threaten to reveal an individual’s identity will be solved according to the recommendations of Tessier and Bonnemains [54].

We do not expect any serious adverse events as a result of using the app. However, we will monitor any negative consequence that results from usage. During the training of the treatment team, we will instruct the health care providers to report negative experiences of the app (eg, frustration due to low digital literacy skills). During biweekly contact with treatment teams, the researcher will also actively ask for any adverse event. In the unlikely event of harm, patients can appeal to the liability insurance of the sponsor that covers any damage to research subjects caused by the study within 4 years after the end of the study.

During all analyses, the treatment condition will be masked to the researchers. We will analyze the data according to the intention-to-treat principle and include patients in the analysis regardless of their adherence to the treatment protocol. We will perform no interim analysis. We will collect data at the following time points: –t₁ (baseline), t₁ (immediately postintervention), t₂ (3 months postintervention), t₃ (6 months postintervention), and t₄ (12 months postintervention). Prior to the main analysis, we will check the randomization by examining the distribution of baseline characteristics between both groups.

Patient eligibility screening, informed consent, and treatment allocation procedures, as well as the baseline assessment, will be completed prior to the start of the treatment program. Patients in the enhanced treatment condition will receive instruction on how to download and use the app, and both strategies will be explained by a member of the treatment staff. To match the existing treatment content and procedures, the moment that the app will be introduced to patients can vary between locations. During the treatment phase, the app is considered a supplement to the main treatment activities and patients are free to decide how and when to use the app. Health care providers will be instructed to encourage and facilitate using the strategies when they expect it will reinforce the treatment process. However, both strategies can largely be used independently of any treatment activity and are expected to minimally affect direct treatment time. A member of the research team will be available to the treatment staff throughout the experimental phase for additional questions, support, and discussions regarding optimal use. Following treatment, there will be no additional monitoring in the enhanced treatment condition. Patients will continue to be able to use the app at their discretion. Posttreatment data will be obtained directly posttreatment (t1), and at 3 months (t2), 6 months (t3), and 12 months (t4) posttreatment.

This study will evaluate the AGRIPPA app in the context of interdisciplinary multimodal pain treatment programs. Specifically, we will investigate the effect of app use on long-term pain disability and efficiency by means of a cost-effectiveness analysis. To discover how patients interact with the app, we will also explore usability and engagement and test the impact of these variables on pain disability. Together, these analyses will help to demonstrate to what extent the AGRIPPA app contributes to preventing relapse in pain-related disability.

In contrast to prevailing intervention development guidelines, the AGRIPPA development project adopted a co-design approach and started with collecting qualitative data from end users (eg, patients and health care providers) rather than with formulating a theoretical framework. Although co-design is increasingly acknowledged in the health care domain as a method to integrate stakeholder input into the intervention design, more robust evaluations of co-design–based interventions are required to determine its additional value to existing development practices [71,72]. A similar point can be made for the evaluation of mHealth apps. A recent systematic review revealed that health care apps to promote self-management in chronic conditions have seldom been evaluated by randomized controlled trials over a prolonged study period [27]. This study will help understand if an mHealth app that has been developed by co-design methods not only contributes to an acceptable and user-friendly intervention but also leads to maintenance of treatment gains for patients with chronic pain.

Because IMPT programs often substantially vary in dose and content, the inclusion of multiple treatment centers will positively influence generalization. Furthermore, this study builds on a feasibility study where evaluations related to form, content, and integration within treatment programs have been incorporated into the current app and study procedures.

The exploratory analysis of engagement variables in this study is expected to provide preliminary insight into patient adherence to the behavior regulation strategies within the app. According to Sieverink and colleagues [73], insight into adherence to mHealth apps can be acquired by combining usage behavior data with a description of intended use and a well-substantiated justification for this intended use. Although this may be difficult to quantify as the intended usage of the app depends on fluctuations of patients’ functional status in the posttreatment phase, the comparison of system usage data with our expectations regarding the use of the strategies will at least provide an indication of adherence to the app. Possible follow-up studies that include qualitative evaluations of patient input may lead to a more sustained insight into adherence.

This study includes several challenges and compromises that can potentially bias the outcomes. First, including patients with both treatment conditions within one center increases the risk of contamination. Second, health care providers have a large influence on participant engagement. Our feasibility study indicated that health care provider involvement varied greatly between patients, and that patients with limited health care provider feedback did not always use the intervention as intended (see personal communication, first author SE; manuscript under review). By scheduling regular contact moments to discuss progress, we aim to minimize the impact of this potential threat. Third, the limited project duration and funding resulted in a maximum follow-up period of 12 months. Although this first year may be crucial for integrating the newly learned management strategies into a daily life routine, the effect of the AGRIPPA app on late-onset occurrences of relapse will not be monitored in this study. Limited funding also prevented the development of an active (mHealth) control condition. Although opting for a treatment-as-usual condition as a comparator is a widely used method, this does not control for the potential placebo effect of receiving an mHealth app. Fourth, to minimize the impact for patients to participate in this study, we have selected a limited number of outcome measures and measurement time points in addition to routine assessments. This may result in an increased recall bias or limited insight in potential factors that could explain a possible effect of the app. Finally, patients will require sufficient digital literacy skills to effectively use the app, which may lead to self-selection during the recruitment phase. The selection bias may threaten generalization.

Maintaining behavior change is notoriously difficult to achieve and every small step toward decreasing relapse or understanding the specific mechanisms by which relapse occurs or is prevented will be important to the field of pain rehabilitation. Implementation of this intervention in treatment programs may also positively empower patients to take a more proactive role in their treatment program and increase sharing their experiences, thoughts, and beliefs with health care providers and significant others. This may not only lead to a better patient–health care provider relationship and improved mutual understanding but is also expected to positively influence adherence to newly learned pain management strategies [74].