This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on http://www.researchprotocols.org, as well as this copyright and license information must be included.
Low back pain (LBP) is a leading cause of disability worldwide. Most patients with LBP encountered in primary care settings have nonspecific LBP, that is, pain with an unknown pathoanatomical cause. Self-management in the form of physical activity and strength and flexibility exercises along with patient education constitute the core components of the management of nonspecific LBP. However, the adherence to a self-management program is challenging for most patients, especially without feedback and reinforcement. Here we outline a protocol for the design and implementation of a decision support system (DSS), selfBACK, to be used by patients themselves to promote self-management of LBP.
The main objective of the selfBACK project is to improve self-management of nonspecific LBP to prevent chronicity, recurrence and pain-related disability. This is achieved by utilizing computer technology to develop personalized self-management plans based on individual patient data.
The decision support is conveyed to patients via a mobile phone app in the form of advice for self-management. Case-based reasoning (CBR), a technology that utilizes knowledge about previous cases along with data about the current patient case, is used to tailor the advice to the current patient, enabling a patient-centered intervention based on what has and has not been successful in previous patient cases. The data source for the CBR system comprises initial patient data collected by a Web-based questionnaire, weekly patient reports (eg, symptom progression), and a physical activity-detecting wristband. The effectiveness of the selfBACK DSS will be evaluated in a multinational, randomized controlled trial (RCT), targeting care-seeking patients with nonspecific LBP. A process evaluation will be carried out as an integral part of the RCT to document the implementation and patient experiences with selfBACK.
The selfBACK project was launched in January 2016 and will run until the end of 2020. The final version of the selfBACK DSS will be completed in 2018. The RCT will commence in February 2019 with pain-related disability at 3 months as the primary outcome. The trial results will be reported according to the CONSORT statement and the extended CONSORT-EHEALTH checklist. Exploitation of the results will be ongoing throughout the project period based on a business plan developed by the selfBACK consortium. Tailored digital support has been proposed as a promising approach to improve self-management of chronic disease. However, tailoring self-management advice according to the needs, motivation, symptoms, and progress of individual patients is a challenging task. Here we outline a protocol for the design and implementation of a stand-alone DSS based on the CBR technology with the potential to improve self-management of nonspecific LBP.
The selfBACK project will provide learning regarding the implementation and effectiveness of an app-based DSS for patients with nonspecific LBP.
RR1-10.2196/9379
The recent Global Burden of Disease Study showed that low back pain (LBP) is the most significant contributor to years lived with disability worldwide [
selfBACK addresses nonspecific LBP, that is, pain with an unknown pathoanatomical cause, which comprises >85% of all patients with LBP observed in primary care settings [
Many patients with long-term conditions find it challenging to self-manage their illness, for example, through lifestyle modifications, with little or no additional support [
In this paper, we outline a comprehensive protocol for the design and implementation of an evidence-based decision support system (DSS), selfBACK, which has the potential to improve self-management of nonspecific LBP. The core of selfBACK is to (1) provide effective evidence-based advice on physical activity and tailored exercise training according to personal goals, personal characteristics, symptom progress, and functional ability and (2) provide educational material to individuals on self-management of their LBP condition. The resulting selfBACK system constitutes a data-driven, predictive DSS that uses the case-based reasoning (CBR) methodology [
The selfBACK project is a Research and Innovation Action funded under the Societal Challenges—Health, Demographic Change, and Well-Being call of the Horizon 2020 program. The project runs from the start of 2016 until the end of 2020. The overall aim of the selfBACK project is to improve self-management of nonspecific LBP to reduce pain-related disability.
The self-management plan will be tailored to each patient according to data collected by a baseline Web-based questionnaire, a weekly question and answer (Q/A) session in the selfBACK app, and a physical activity-detecting wristband worn by patients. Besides developing an infrastructure for collecting and processing these data, objective 1 also includes work that focuses on the definition of case representations and similarity measures, which are the core components of the CBR technology.
Specifically, the selfBACK system is designed to assist patients in deciding upon and reinforcing the appropriate actions to manage their LBP. Based on the current best evidence, specific content for supporting physical activity, patient education, and strength and flexibility exercises are developed as part of the DSS. Besides CBR for handling situation-specific knowledge, elements of the model- and rule-based reasoning are used to capture and utilize generalized knowledge (eg, clinical guidelines) as well as customize the recommendations for self-management.
Overview showing how the overall aim is achieved by the selfBACK objectives. DSS: decision support system; LBP: low back pain.
The decision support is conveyed to patients by a mobile phone app. The app provides patients with (1) instant feedback on the activity level and activity distribution (based on the data stream from the wristband) in accordance with the personal goals set by them and (2) tailored educational sessions and specific exercise training in line with patients’ goals, personal characteristics, symptom progress, and functional ability.
The effectiveness of the selfBACK app will be evaluated in a multinational RCT (parallel group trial) that will target care-seeking patients with nonspecific LBP. The comparator will be patients who receive the usual treatment. In addition, a process evaluation will be carried out as an integrated part of the RCT to document barriers and facilitators for the uptake and utilization of the selfBACK app and for gaining an understanding of patient experiences of using the app. Furthermore, a detailed protocol of the RCT will be reported in a separate publication.
The concept underlying the design of the selfBACK system is that an improved clinical outcome in patients with LBP will rely on behavioral change, which, in turn, might be conditioned by several factors, such as fear avoidance, pain self-efficacy, comorbidities, mood, and physical and mental capacity. The literature offers numerous health behavioral change theories that explain and predict the physical activity behavior, several of which focus on the motivation and volition, in other words, intending to be active and transforming the intention into action [
In the following section, we describe the technical solution that will be implemented to address the project objectives and concept underlying the selfBACK DSS.
The core of selfBACK is a DSS that helps patients to follow a plan for physical activity (ie, daily step count), education, strength, and flexibility exercises according to personal goals, personal characteristics, symptom progress, and functional ability. To accomplish this, selfBACK incorporates existing knowledge (eg, clinical guidelines and medical ontologies) and information provided by patients to recommend tailored advice for self-management.
The decision support is conveyed to patients via a mobile phone app in the form of advice for self-management. The app will be developed for Android and iOS using the React Native framework. The process for producing and tailoring the self-management plans and is illustrated by steps 1-6 in
Illustration of the overall architecture and how the data processing of the person-directed modules link together in the selfBACK system.
Components of a case description within the case-based reasoning (CBR) system containing the patient characteristics and the compiled advice. Only a relevant sub-set of the weekly questions will be asked in each session. The questionnaires used to collect baseline information are described in
For example, reporting of poor sleep will initiate an educational session targeting sleep behavior accompanied by appropriate follow-up questions. In contrast, reporting of good sleep will reduce the frequency of asking about sleep. Moreover, the regular monitoring of factors such as fear avoidance and pain self-efficacy by the weekly Q/A sessions allows us to make decisions about the amount and type of feedback and educational guidance required by individual users of the selfBACK app. For example, patients with low fear avoidance and high pain self-efficacy may only require simple educational follow-up information provision, whereas patients with high fear avoidance and low pain self-efficacy would require more support and guidance on behavioral change techniques, such as goal setting, pacing, and action planning. In addition, the activity data (3) provide information about whether patients follow the suggestions in the self-management plan (such as the number of steps per day). Regarding the baseline and Q/A data, activity log data are also sent to the selfBACK server giving input to the periodic run of the CBR cycle to generate an adjusted plan for self-management. Furthermore, this means that patients will receive a general advice on physical activity during the first week and that the goal for the number of steps per day will be adjusted in the consecutive weeks according to the incoming data from the wearable.
In addition to the follow-up and feedback on the daily step counts, the selfBACK system provides patients with recommendations on specific strength and flexibility exercises. The exercise plan conveyed to patients includes instructions and illustrations of the exercises along with recommendations for the number of sessions per week and repetitions or sets within a session. A typical exercise plan is designed to target strengthening of the back extensors, abdominals, gluteal muscles, and core muscles along with the flexibility of knee and hip and trunk muscles with short videos demonstrating the execution of the exercises. At present, insufficient evidence exists to make strong recommendations for or against any specific strength or flexibility exercises [
The selfBACK system constitutes a data-driven, predictive DSS that uses the CBR methodology to capture and reuse patient cases to suggest the most suitable self-management plan for an individual patient [
Overview of the information collected at the baseline.
Baseline information | Response options or questionnaire reference | ||
Demographics | Age, gender, height, weight, family, ethnicity, education, employment status | ||
Physical work characteristics | Saltin–Grimby Physical Activity Scale [ |
||
Average last week | Visual analog scale, 0-10 [ |
||
Worst last week | Visual analog scale, 0-10 [ |
||
Length of current episode | <1 week; 1-4 weeks; 5-12 weeks; >12 weeks | ||
Days with LBP past year | 0 days; 1-7 days; 8-30 days; >30 but not every day; every day | ||
Use of pain medication last week (days) | None; 1-2 days; 3-5 days, daily | ||
Pain-related disability | Roland-Morris Disability Questionnaire [ |
||
Function | Patient Specific Functional Scale [ |
||
Reduced work activity | Yes; no | ||
Reduced leisure time activity | Yes; no | ||
Current work ability | Work ability index, 0-10 (single item) [ |
||
Leisure time physical activity | Saltin–Grimby Physical Activity Scale [ |
||
Comorbidities, musculoskeletal | Pain mannequin | ||
Comorbidities, others | Cardiovascular disease; heart failure; stroke or brain hemorrhage; asthma; chronic bronchitis or emphysema, COPDb; diabetes; gastrointestinal problems; kidney disease; cancer; epilepsy; osteoporosis; osteoarthritis; depression; anxiety; sleep apnea; rheumatoid arthritis; psoriatic arthritis or psoriasis; other | ||
Quality of life | EQ-5Dc [ |
||
Sleep problems | Sleep Screening Questionnaire [ |
||
Fear-avoidance beliefs | Fear-Avoidance Beliefs Questionnaire [ |
||
Pain self-efficacy | Pain-Related Self-Efficacy Questionnaire [ |
||
Illness perception | Brief Illness Perception Questionnaire [ |
||
Perceived stress | Perceived Stress Scale [ |
||
Mood | Patient Health Questionnaire [ |
aLBP: low back pain.
bCOPD: chronic obstructive pulmonary disease.
cEQ-5D: European Quality of Life-5 Dimensions.
The selfBACK architecture consists of clients that are connected to a server, which holds the DSS. The architecture covers the following two main scenarios. First, a patient is provided with the necessary credentials to access a web interface to be used for the initial sign-up and to fill out a baseline questionnaire, and second, a patient is equipped with a wearable and the selfBACK mobile phone app, which guides self-management according to the goals set by the patient. The mobile phone app synchronizes with the wearable and obtains the patient’s activity log and sends push notifications with content that encourages physical activity. Furthermore, the mobile phone app itself is the tool for the patient to obtain personalized information and educational explanations.
The decision support server performs all relevant tasks for maintaining the knowledge repository and provides the infrastructure for the advice generation services. Those services are parts of the server and communicate with clients through a secure access layer. The data generated will be used for updating each patient’s self-management plan.
The clients are either native mobile phone app or web browser users accessing the selfBACK system. The server can only be accessed through the secure access layer, which requests authentication and maintains user groups and secure workflows ensuring only relevant data are accessible to any client. In addition, the decision support server performs all relevant tasks from preprocessing incoming data and creating abstractions to running the decision support engine; this process is enhanced by machine learning tasks, which enable the creation of new cases and rules from incoming data. The knowledge obtained from this process is stored permanently in the knowledge repository.
The core of the DSS is the CBR system for finding the most similar patient cases. An additional rule-based system module captures generalized knowledge that is complementary to the situation-specific knowledge in cases in the form of clinical guidelines and generalizations over cases. Depending on its type, each incoming data stream is preprocessed in a particular way. In addition, data abstraction enhances the incoming data with domain knowledge, thereby allowing more comprehensive reasoning to support the decision making. Furthermore, implementing case matching as software services allows selfBACK to scale and perform the generation of self-management plans in parallel, whereas the core engine and knowledge repository stay consistent.
The knowledge repository holds cases and rules as well as the underlying ontology, and provides this knowledge for the DSS. The case base, containing cases, is the primary source of knowledge. Rules provide additional knowledge, particularly targeted for representing relevant clinical guidelines and knowledge. Furthermore, the ontology defines the concepts used by cases and rules with essential relations such as the taxonomical relations that enable the inheritance inference within the ontology.
Preprocessing is performed on all data sources, which are included in the target knowledge models. Data are structured and grouped according to the domain and information types using the ontology. In addition, incoming data are cleaned, normalized, and transformed, and the significant features and instances are selected. The preprocessing strategy depends on the source data and its purpose in the knowledge model.
Data abstraction is applied to make data streams comparable and prepare data for the next process, the personalized decision support. selfBACK uses existing state-of-the-art methods for detecting trends in raw data. In addition, the individual activity streams are processed to reduce complexity or enhance them with the information required for better matching. Once data are delivered to the DSS, each patient is represented as a self-management agent who uses the incoming data to build up a query (a case with a problem description only) and match it against the case base. Consequently, it receives the best matching case that is subsequently personalized to give appropriate advice, enhanced with explanations. Then, an updated self-management plan is returned to the user. Explanations might be justifications for the advice, that is, “how was this advice derived?,” and an explanation of the effects of the advice given the current situations of patients. Furthermore, explanations are stored as predefined text elements in a separate part of the knowledge repository.
In this study, structured intervention mapping is used to guide the development of the content for the selfBACK system [
Throughout the development of the selfBACK DSS, patients with LBP and health care professionals have been interviewed and asked about their experience with the traditional treatment of back pain and how they usually self-manage their LBP. In addition, a panel consisting of clinicians (eg, physiotherapists, chiropractors, sports physiologists, and psychologists) provided feedback and answered a survey concerning the choice of physical exercises and the educational content. Furthermore, the selfBACK team members (eg, physiotherapists, chiropractors, exercise physiologists, and medical doctors) contributed to group discussions and the structuring of the content implemented in the selfBACK DSS. Finally, the developmental versions of the selfBACK app are continuously tested by patients and team members in iterative rounds, during which information is collected in interviews and group sessions with potential users.
The selfBACK project was launched in January 2016 and will run until the end of 2020. The version of the selfBACK DSS that will be used in the RCT will be completed in the fall of 2018. Our target population is care-seeking patients in primary care settings diagnosed with nonspecific LBP. Clinicians will identify patients who will be eligible for self-management and participation in the RCT. The recruitment of patients and data collection in the RCT will start in February 2019 with pain-related disability at 3 months as the primary outcome with additional follow-ups at 6 weeks, 6 months, and 9 months. In addition, results for the trial will be reported according to the CONSORT statement [
Nonspecific LBP is a condition with large interindividual variation in symptoms, treatment responses, and outcomes and is, therefore, suitable for personalized care. Recent studies have shown that a stratified care approach for LBP results in a substantially better treatment response compared with treatment as usual [
In selfBACK, we will collect data about patients by a baseline Web-based questionnaire, an activity-detecting wearable, and weekly Q/A sessions in the selfBACK app. The obtained information will be used to personalize advice to optimize and reinforce self-management of nonspecific LBP. In addition, the collected data, along with the personalized plans for self-management, will be used to build and add new cases to the system’s knowledge repository. The system’s automatic learning component enables new knowledge and plans for self-management to be integrated into the system’s knowledge repository, whereas experiential case-based learning enables improved patient support over time. Therefore, the selfBACK system is a powerful tool to facilitate, improve, and reinforce self-management of nonspecific LBP. Furthermore, the use of the selfBACK app does not require direct medical supervision and can easily be made available to a large number of people implying a cost-effective use of resources. However, selfBACK is not intended to replace clinical care, and we will adhere to the Health On the Net Foundation (HONcode) principle [
The educational and supportive material, general physical activity, and specific strength and flexibility exercises constitute the main components of LBP self-management. In selfBACK, the patients’ physical activity level is monitored and followed-up by data provided by an activity-detecting wearable. The wearable stores data for future synchronization and users do not need to have the phone connected at all times. The monitoring of physical activity will be accompanied by motivational textual feedback displayed on the mobile phone screen. Based on the available sensors embedded in the wearable (eg, three-axis accelerometer), different health metrics are derived, including the number of steps taken and the duration of inactivity. Moreover, we can derive the time a patient is not wearing the wearable to indicate the user pattern of the selfBACK system. Although recent studies have indicated that wearables have poor validity in estimating the energy expenditure [
Many patients find it challenging to self-manage their illness with little or no additional support and the adherence to self-management programs is commonly poor [
A strength of the selfBACK approach is the strong theoretical underpinning along with the use of data-driven computer modeling to tailor advice and follow-up on self-management. Furthermore, because the advice is grounded in the system’s growing experience on the effect of plans for self-management and the accompanying symptom progression, the prediction quality of selfBACK will increase over time. Thus, the selfBACK app could potentially become a potent tool for supporting self-management in patients with LBP. Nevertheless, the risk of poor engagement or the lack of sustained participation exists. Importantly, a process evaluation will be conducted along with the RCT, enabling us to document the implementation of selfBACK, including patient experiences of using the app. This will provide clues about how patient-centered DSSs, such as selfBACK, should be designed to maximize uptake and utilization.
By adapting and advancing the state-of-the-art technology in data capture, data analysis, and proactive decision support, we will, in the selfBACK project, develop and document a DSS to be used by patients themselves to support self-management of nonspecific LBP. In line with current evidence-based recommendations and guidelines, selfBACK incorporates physical activity, education, and specific strength and flexibility exercises to improve, facilitate, and reinforce the self-management process. Furthermore, the effectiveness of selfBACK will be evaluated in a multinational RCT, targeting care-seeking patients in a primary care setting.
case-based reasoning
consolidated standards of reporting trials
consolidated standards of reporting trials statement for randomized controlled trials of electronic and mobile health applications and online telehealth
chronic obstructive pulmonary disease
decision support system
European Quality of Life-5 Dimensions
Health On the Net Foundation
low back pain
randomized controlled trial
question/answer
The selfBACK project has received funding from the European Union Horizon 2020 research and innovation program under grant agreement No. 689043.
None declared.
Evaluation Summary Report from the European Commission.