Chronic low back pain (CLBP) is a major public health concern worldwide because of its economic and social consequences [1,2]. It is the leading cause of disability and work absenteeism, and its prevalence is still increasing [3,4]. Management of daily symptoms and activity are major challenges for people with this condition [5-7].

Self-management strategies and maintaining an active lifestyle have been consistently recommended for people with CLBP [8-12]. Self-management can be defined as the ongoing and dynamic ability to handle symptoms, such as pain, physical and psychological consequences, and lifestyle adjustments [13,14]. Self-management programs include educational and psychosocial interventions; maintaining an active lifestyle is a core component of these programs. People with CLBP are encouraged to engage in regular physical activity and to adopt healthy behaviors to manage their symptoms [14]. Although self-management and physical activity–based interventions improve pain [15] and reduce disability [16,17] in the long term, as well as promote the development of self-management skills [18,19], the effect of such interventions are small to moderate. Moreover, it is difficult to support self-management in clinical practice because of constraints in time, service organization, and follow-up [20]. Therefore, there is a need to identify innovative interventions to promote sustainable self-managed physical activity in people with CLBP.

Recent advances in technology like mobile health (mHealth) may offer new opportunities to support self-managed physical activity and pain in people with CLBP. mHealth is defined as a “medical and public health practice supported by mobile devices, such as mobile phones, patient monitoring devices, personal digital assistants, and other wireless devices” [21]. It can help people change their health behaviors [22] and help people with CLBP pursue specific physical activity goals while receiving continuous feedback on their physical performance [23,24]. It could be used to help them manage their condition and maintain an active lifestyle [25,26]. From the point of view of health care professionals, mHealth technology can provide individualized interventions with real-time feedback, support their coaching role [27], and support the development of behavior change [28,29].

Although mHealth technologies are promising, their content and the context in which they are offered need to be explored further. Many mobile apps are available to support people with CLBP in self-management, mainly providing exercise recommendations or information about the mechanisms of CLBP. These apps score poorly on the Mobile Application Rating Scale (a scale to assess app quality), and do not mention the theoretical approaches on which they are based [30-32]. Sometimes these mobile apps are combined with a physical activity tracker; however, they have shown limited effectiveness in disability and pain management in people with CLBP [33,34]. For example, several studies using the Fitbit device found nonsignificant results for pain [35,36] and disability [37]. This lack of an effect could relate to the fact that the target population for the tool is healthy individuals. Devices, such as the Fitbit were developed primarily to help young people improve their physical condition and they do not specifically consider people with CLBP and their context. They do not appear to be suitable for supporting health care professionals or people with CLBP who may benefit from increasing their participation in unstructured physical activity, like walking, and reducing sedentary lifestyle habits [38-41]. Moreover, studies of rehabilitation programs that integrated mobile technology found mixed results because of participants’ lack of adherence to the wearable technology [35,36].

It is now widely recognized that user-centered design (UCD) approaches are needed to facilitate the development, acceptability, and implementation of mHealth technologies [42-45]. In the field of CLBP, such designs are seldom used and rarely rigorously applied [46], particularly in the initial stages of device design [47].

The main aim of this study was to explore the needs of people with CLBP and health care professionals regarding the use of mHealth technologies to support self-managed physical activity and pain. The specific aims were to (1) investigate how technology could help people with CLBP overcome the difficulties of managing their own physical activity with an expert group, (2) explore the needs, experiences, and preferences of people with CLBP and health care professionals regarding technology to support self-managed physical activity, and (3) validate the identified needs and develop a set of mHealth technology design recommendations.

A participatory study was conducted using a UCD approach. UCD is an iterative process that focuses on users and their contexts in all stages of development design [44,48,49]. This study focused only on the initial process of the approach to investigate the needs of people with CLBP and to guide future prototype designs (Figure 1) [50-52]. The study was conducted in France and Switzerland and involved three successive data collection phases: (1) needs investigation with a group of experts, (2) needs exploration with people with CLBP and health care professionals, and (3) validation of the needs identified and development of recommendations. It was conducted after a preliminary literature review that aimed to identify existing evidence on mHealth devices used to support self-managed physical activity in people with CLBP.

The study was approved by the research ethics committee of the University of Lyon (CER-UDL 2022-10-13-003). The Cantonal Ethics Committee of Swissethics stipulated that this research was outside their jurisdiction and did not require ethics approval (Req-2022-00733).

Descriptive statistics were used to characterize the participants in the 3 phases, using frequencies and percentages. In phase 1, participant discussions were recorded and transcribed verbatim for analysis using an inductive thematic analysis [53,54]. First, 2 researchers (MB and AD-B) independently read the transcripts to familiarize themselves with the content. Second, they separately coded the data using NVivo (version 20, Lumivero) to create an initial codebook. After the initial coding, the codebook was discussed between the 2 researchers, and segments of the content with similar meaning were assigned to the same code. Third, the coded transcripts were used to refine the concepts of the initial codebook and combine the codes into key themes. When new codes or themes emerged, the codebook was revised, and the previous transcripts were recoded. Any discrepancies between the researchers were resolved by discussion, and any necessary adjustments were made. This thematic analysis was then conducted again to reach a consensus. Finally, the content of the themes and subthemes of the coding scheme was discussed with the entire research group.

In phase 2, the data analysis procedure from phase 1 was repeated using recorded material from all the focus groups [53,54] and the field notes written by the moderator in charge of describing the group interactions. The final version of the codebook from phase 1 was reused and revised by 2 researchers for use in phase 2 (MB and AD-B). All themes were validated by the entire research group.

In phase 3, the level of expert consensus was calculated for each item after each round [55]. Consensus was considered achieved when 90% of the participants agreed or strongly agreed with a statement. All suggestions were reviewed by 2 researchers (MB and AD-B), and the data were coded using the same codebook used for phases 1 and 2.

A total of 121 participants were involved in the study. Of the 121 participants, 9 experts took part in the first phase, including 4 subgroups of 2 to 3 experts in the first part of the workshop and 2 subgroups of 4 to 5 experts in the second part of the workshop (one of the subgroups was composed of 1 patient who was an expert, 1 clinician, 1 rehabilitation researcher, and 1 biomechanical engineer while the other was composed of 1 patient who was an expert, 2 clinicians, 1 rehabilitation researcher, and 1 biomechanical engineer). Of 26 participants who took part in the second phase, 11 people with CLBP and 15 health care professionals (including 4 and 7 participants, respectively, for the 2 sets of focus groups with people with CLBP, and 7 and 8 participants, respectively, for the 2 sets of focus groups with health care professionals). Of 86 participants who took part in the third phase, 45 people with CLBP and 41 health care professionals. About 63% (7/11) of the participants with CLBP and 46% (7/15) of health care professionals from phase 2, participated in phase 3. The demographic characteristics of the participants are presented in Table 1 and more details about the participants are available in Multimedia Appendix 1.

More than 12 hours of participatory design workshop data, focus group data, and suggestions from 2 questionnaires were coded into 5 themes that are described in more detail below: (1) difficulties experienced by people with CLBP in relation to self-managed physical activity and pain, (2) device concept (aim and strategies), (3) device content, (4) condition of use, and (5) facilitators and barriers to use.

The difficulties faced by people with CLBP in relation to self-managed physical activity and pain were specifically explored in phase 1 by the expert group composed of patients who were experts, clinicians, rehabilitation researchers, and biomechanical engineers. All participants were asked to discuss these difficulties to reach a consensus on the problems faced by people with CLBP. These difficulties were then iteratively addressed by people with CLBP and health care professionals in phases 2 and 3. These difficulties are presented in Table 2. The more prevalent ones included pain, personal beliefs, difficulty pacing physical activity, and a lack of sustained, individualized, long-term support from health care professionals. Only these prevalent difficulties were used as the basis for the assessment of the needs of people with CLBP and health care professionals, as they were shared by both groups.

The needs of people with CLBP and health care professionals regarding the device concept are described in Tables 3 and 4 and illustrated with verbatim. All participants highlighted the importance of having a device that would not only help increase physical activity level of people with CLBP but also better pace it. They insisted on the need for this device to also help people with CLBP with long-term pain management. To achieve these objectives, they suggested the device should support the following behavior change strategies: goal setting, self-monitoring, feedback, positive reinforcement, reward delivery, and social support.

The needs of people with CLBP and health care professionals regarding device content are described in Tables 5 and 6 and illustrated with verbatim. Participants expressed the need for a device that would allow people with CLBP to automatically track the following data: number of steps, heart rate, activity versus rest time, intensity of the activity, stress, and sleep. They emphasized the need for the collection of additional data recorded by the user, such as an activity diary, well-being, pleasure, satisfaction, comfort, medications taken, pain, sleep, and stress. They insisted that the user should specifically select automatically-collected and self-reported data. Participants also wanted the device to be able to send notifications to help people with CLBP manage their physical activity and pain. They highlighted the usefulness of receiving regular activity reports and of being able to access specific resources, such as physical exercises, relaxation exercises, and questionnaires to refresh knowledge about pain. They suggested the use of a wristband combined with a digital app that is easy to use, discreet, comfortable, waterproof, and robust.

The needs of people with CLBP and health care professionals regarding the conditions of use are described in Table 7 and illustrated with verbatim. Participants indicated that this device may be used either 24 hours per day or only during the day, depending on the user’s preference. They wanted it to be provided by a health care professional as part of a rehabilitation program, and they wanted health care professional supervision for its use. They wanted to be able to consult the data collected in real time, but some health care professionals expressed concern about misinterpretation and suggested that the data should initially be only available in the presence of a therapist. All participants insisted on the importance of long-term follow-up by a health care professional to check the use of the device and to help adjust the goals (either face-to-face or by videoconference).

The facilitators and barriers to the use of a device mentioned by the participants in all the study phases are described in Table 8. The facilitators mentioned corresponded to the functionalities and features considered necessary for the content of the device. The potential barriers were related to concerns about the validity of the data collected, the storage and confidentiality of that data, the need to recharge the device, the additional availability of health care professionals to provide and monitor the device, and the risk of adverse effects associated with the use of the device.

The needs of people with CLBP and health care professionals regarding device concept and content and the conditions of use were confirmed in 2 rounds of Delphi questionnaires with 86 and 61 participants, respectively (Tables 9 and 10). Twenty-eight needs were selected from those identified in the previous 2 study phases. A lot of attention was given to the variables that could be collected or recorded by the device, as much of the discussion in phases 1 and 2 focused on these aspects. In the first round, consensus was reached on 13 (46%) of the 28 needs (the 13 validated needs are italicized in Table 9). In the second round, the 15 needs for which there was no consensus were modified based on the suggestions made in the first round of consultation: 11 (73%) of the 15 needs were reformulated, and 3 (20%) of the 15 needs were merged into a single need (changes in needs are italicized in Table 10). In the second round, consensus was reached on 11 (91%) of the 12 needs. Out of 25 needs in total, a consensus was reached on 24 (96%) needs.

The needs validated by the participants were synthesized as design recommendations using the Behavioral Intervention Technology model and are illustrated in Figure 2. This is based on the illustration of the MyFitnessPal mHealth app by Mohr et al [56] and describes the expected functionalities and technical features of the device. The numbers shown illustrate the conditions of use and the expected workflow: (1) first, the automatic data collected by the device can be selected, then (2) the data collected by the user can also be selected. From this, (3) an activity objective can be defined, (4) and the user can follow changes in the values of variables collected or recorded, referring, if he or she wishes, to the physical or relaxation exercises proposed. (5) Reception of regular notifications and alarms can be set.

This study explored the needs of people with CLBP and health care professionals regarding the use of mHealth technology to support self-managed physical activity and pain. It elucidated how people with CLBP believed the adoption of an mHealth device could mitigate their difficulties in managing physical activity in relation to their pain, beliefs, difficulties in pacing physical activity, and the need for long-term individualized support.

The results of this study show that participants (people with CLBP and health care professionals) expressed interest in using an mHealth system to help people with CLBP manage their physical activity and pain over the long term. They wanted a device that would allow them to set personalized activity goals, monitor their activity, receive feedback on their performance, receive positive reinforcement, possibly receive rewards, learn information, and receive social support. These results provide a more precise definition of the needs identified by Merolli et al [46], who highlighted the usefulness of developing a technological solution to support the self-management of CLBP with tracking, notifications, feedback, provision of educational resources, and exercises. These results, collected using a rigorous process, accurately describe the expectations of end users in terms of the characteristics of a potential device.

Special attention was paid to the functionalities and variables recorded by the device. This is a fundamental step in the development of effective technology [57]. People with CLBP and health care professionals described the need to set realistic activity goals and to track them using an mHealth device, while also having complementary tools to help them manage their pain daily (eg, a physical activity exercise bank and relaxation exercises). They strongly emphasized that physical activity should be monitored using self-selected variables, stating that the need was not always to increase physical activity level, but rather to better pace it daily through achievable goals. Several participants with CLBP highlighted their tendency to overdo activity, debunking the assumption that physical activity levels in people with CLBP are often lower than in healthy people [58-60].

Overall, people with CLBP and health care professionals expressed similar needs in terms of an mHealth device. They strongly emphasized that the settings should be personalized, and the device should be easy to use. However, the opinions of these groups differed regarding the variables that should be collected and the components that should be provided by the device. Health care professionals believed the focus should not be on the assessment of negative variables, such as pain, whereas people with CLBP wanted to collect this type of variable. Health care professionals suggested that a “challenge” module should be included, with a system to collect points or badges, to encourage users to interact with the device. The opinions of the people with CLBP were more divided about such a module, with some very interested in this type of reward and others completely opposed to it because they were afraid of being involved in competitive challenges. The gamification of mHealth is attracting increasing interest as a means of encouraging changes in physical activity behavior, although current evidence on the effectiveness of this type of device is still limited [61,62].

The potential disadvantages and adverse effects associated with the use of a device to self-manage physical activity were discussed in detail during the 3 phases of the study. On the one hand, the effects of such a device on the development of new behaviors and lifestyles may be directly linked to its maintenance. For example, behavior change is less likely to occur if the device monitor is lost, broken, not charged, or forgotten. In contrast, if the device is used consistently and integrated into an intervention aimed at maintaining new behaviors, new habits could emerge in the medium term [63,64]. In addition, some health care professionals were concerned that people with CLBP could misinterpret the data collected by the device, and that this could induce feelings of guilt. Indeed, some health care professionals were against users being able to consult their data in real time, whereas people with CLBP saw this as an obvious possibility. The ability of the mHealth device to provide feedback to the user is considered important for encouraging behavioral change [61,65]. Health care professionals also mentioned the additional time required to teach their patients to use the device. In addition, both health care professionals and people with CLBP expressed concern about the validity of automatically-collected data, its security, and the individual’s privacy.

The discussions went well beyond the functionality of the device. The results highlighted the importance of developing an intervention around the mHealth device to help people with CLBP manage their activity and pain, rather than simply relying on the mHealth device. Human support was stated as an essential component of long-term care for people with CLBP. These results confirmed the findings of Svendsen et al [66] that technology alone was not enough to encourage people to engage in self-management. All participants were very clear about the conditions in which the device should be made available, that is, its implementation should be supported by a health care professional, and education should be provided on the importance of physical activity, its role in pain management, and the operation of the device. Education is important to increase the effectiveness of device use because people’s beliefs determine their behavior [67]. These findings go beyond those of the pilot study by Ellingson et al [68], which showed the promising effects of a device-based intervention combined with minimal human support for people with CLBP; because they specified the expected characteristics of human support, in particular, the context in which the device should be offered, how it should be supported by health care professionals, and the frequency at which health care professional support should be provided.

After validating the needs of people with CLBP and health care professionals regarding mHealth devices, we developed design recommendations using the Behavioral Intervention Technology model (Figure 2) [56]. This model helped to structure the needs identified during the 3 phases of the study and to link them to the behavior change strategies that could be integrated within a new intervention encompassing the use of an mHealth device. These design recommendations should now be further developed in the light of behavior change theories and taxonomies currently published in the literature (eg, Michie’s taxonomy) [14,69,70]. Future research should also seek to discuss and refine the prototype development phase with end users. Indeed, after this initial conceptualization phase, this synthesis could be revisited at a later stage to continue the prototype development phase, still using a user-centered approach.

This study was based on the experiences of experts and potential end users. It first identified the needs of 9 experts in phase 1, followed by 26 people with CLBP and health care professionals in phase 2, and then validated these needs with a large number of participants (86 in the first round and 62 in the second round). Despite the sample size, the first limitation concerns the population studied, and its impact on the representativeness and transferability of the results. The participants with CLBP and the health care professionals were all involved in multidisciplinary rehabilitation; therefore, the results may not be generalizable to other settings. This choice was made to have a comparable, homogeneous sample, but the needs and experiences of people with CLBP who do not have access to multidisciplinary rehabilitation (ie, only physiotherapy sessions) may be different.

The Delphi method used in phase 3 may have been subject to some bias. Although the needs for device design were developed during phases 1 and 2, not all the elements that emerged in the first 2 phases were specifically examined during phase 3. The elements selected by the research team for the Delphi questionnaire are those that have been the subject of significant debate among the participants in the previous phases. Thus, in phase 3, particular attention was paid to the variables collected by the device, the content, and the condition of use. This process may have left out elements that would have been also important to present in the Delphi questionnaire. In addition, participants involved in phase 2 were invited to take part in phase 3, which might have introduced bias because of their previous involvement in defining the needs. This involvement may have resulted in an overestimation of the consensus level achieved in the 2 rounds of Delphi questionnaires.

This study used a UCD approach to explore the needs of people with CLBP and health care professionals regarding the use of a device to support self-managed physical activity. It identified potential users’ expectations regarding the device’s objectives, behavior-change strategies supported by the device, the device’s functionalities, its technical features, and facilitators and barriers that may influence its implementation in a clinical context. The results were used to delineate design recommendations for the development of an mHealth device specifically for people with CLBP and health care professionals. These recommendations will be used in the future to prototype innovative devices that could be offered by health care professionals to people with CLBP in a follow-up rehabilitation context.