This study is grounded in a constructivist epistemology, in which knowledge is understood as coconstructed through social interaction between participants and researchers rather than discovered as an objective truth. The dataset was generated through repeated semistructured interviews and participant-generated photos [23], and a reflexive thematic approach was applied in the data analysis [24,25].

The authors form an interdisciplinary team of 4 women and 1 man, all based in academic settings at Swedish universities. Four of the authors also have experience in clinical practice within physiotherapy (CKH, DMC, and HL) or occupational therapy (LH). All are experienced in intervention and implementation research, including mHealth (CKH and DMC) and knowledge of behavior change theories (HL and LK). In line with constructivism and reflexive thematic analysis [26], themes are viewed as actively generated through an interpretative and reflexive analytic process that draws on participants’ accounts and researchers’ positionality. Participants with experience of the mHealth version of the i-REBOUND intervention were integral to the knowledge production process, and we acknowledged our active role as researchers in shaping the analysis. The study’s reporting follows the Reflexive Thematic Analysis Reporting Guidelines [27].

We strategically recruited 12 participants (10 women, 2 men) from the experimental group (n=57) of a feasibility randomized controlled trial (n=114) after participating in the mHealth version of the i-REBOUND intervention for people poststroke and TIA [21]. Inclusion criteria for the feasibility randomized controlled trial were adults who (1) experienced a stroke or TIA between 3 months and 10 years prior to the study, (2) lived at home, (3) had access to a reliable internet connection, (4) could walk short distances indoors with or without a walking device, (5) could use a smartphone, and (6) could self-identify through a Swedish secure self-authentication app [28]. Exclusion criteria were (1) currently meeting recommended physical activity levels (ie, at least 150 minutes per week of moderate physical activity or at least 75 minutes per week of vigorous intensity physical activity [29]) or had (2) severe health conditions and/or (3) significant cognitive impairments [28]. Physical activity levels of the participants were assessed via a telephone interview, and individuals were eligible for inclusion if they self-reported not meeting the recommended 150 minutes of moderate-intensity physical activity per week based on frequency, intensity, and duration. The strategic sampling aimed to achieve diversity in age, sex, physical activity level, geography, and stroke or TIA diagnosis. Twelve participants were identified and agreed to participate. Among the recruited participants, 10 regions across Sweden were represented, and the participants were aged 56 years through 80 years at the start of the intervention. Participant characteristics are presented in Table 1.

For full intervention details, see the study protocol [28]. The mHealth version of the i-REBOUND intervention aimed to support home-based exercise and physical activity for individuals poststroke or TIA. The intervention was delivered through an mHealth app called the STAAR (Stroke Treatment through Active and Accessible Rehabilitation) app. During the 6-month intervention, participants received digital supervised support from a physical therapist and had access to self-managed mHealth support. The participants attended weekly exercise sessions of moderate to vigorous intensity (months 1 to 3: two 25-minutes sessions per week; months 4 to 6: one 25-minute session per week) and monthly individual 30-minute counseling sessions led by a physical therapist. The counseling sessions focused on setting physical activity goals, discussing progression, and developing self-management strategies to support maintained physical activity engagement. Examples of goals were “Gym training twice a week on Mondays and Wednesdays. 20 minutes of aerobic exercise and 40 minutes of strength training. Borg scale 13‐14.” and “Brisk walks two to three times per week, 30 to 60 minutes per session.” To further support the promotion of physical activity, the app provided self-managed mHealth support, including educational videos, an activity diary for self-monitoring, and prerecorded exercise videos. During the intervention, a chat function enabled contact with the physical therapist. An action plan, including long-term goals, for maintaining engagement in physical activity postintervention was formulated with support of the physical therapist at the final counseling session. After the intervention, participants continued to have access to self-managed mHealth support, although no further contact with the physical therapist was provided.

Semistructured interviews were conducted in Swedish at 3 months and 6 months postintervention. Interviews were carried out via digital video calls by the first author, a PhD candidate and experienced geriatric physical therapist, between June 2023 and November 2023. In the 6-month interview, participant-generated photo elicitation was used [23]. The first author had brief contact with the participants prior to the interviews, due to her involvement in the data collection in the feasibility randomized controlled trial.

The first and last authors developed and pilot-tested the interview guides. The 3-month interview covered the following conceptual areas: (1) the participants’ perceptions of maintaining physical activity after the intervention, (2) barriers and facilitators to maintaining physical activity after the intervention, and (3) expectations and concerns about maintaining physical activity after the intervention. Examples of interview questions were “What has been your experience of being physically active after the intervention ended?” and “Can you tell me what facilitated regular physical activity for you after the intervention ended?” Before the 6-month interview, the participants were encouraged to take 3 to 5 photos reflecting how they perceived maintaining their physical activity after completing the mHealth version of the i-REBOUND intervention. After a brief reflection on their current physical activity level, the participants were asked to reflect on one photo at a time. Open-ended questions were inspired by Green et al [33] and included “Tell me about this photo—why did you choose to bring this photo?” “Tell me what could prevent you from doing this activity regularly?” and “Was there anything missing in the intervention to make this activity regular?” To enhance the richness of the descriptions, participants were prompted with follow-up questions and encouraged to elaborate on their responses using examples of situations. The interviews lasted between 40 minutes and 90 minutes and were audio-recorded. Field notes were written after each interview. Interviews were transcribed and anonymized before the data analysis.

Analysis followed reflexive thematic analysis, based on the 6-phase process outlined by Braun and Clark [24,25]. The primary focus was on the transcribed text; however, photos were analyzed alongside participants’ interviews to add context and generate a richer understanding of the participants’ verbal accounts [34]. The software program N-Vivo 15 was used to organize the transcribed data. In phase one, familiarization with data, the first author familiarized herself with the dataset by carefully reading and listening to all interviews while simultaneously taking field notes. In this phase, discussions were continuously held with the last author, who read a sample of the interviews. In phase two, generating initial codes, the first author, in collaboration with the last author, identified and labeled parts of the text that captured meaningful experiences related to the research aim. The third phase, generating initial themes, included reflexive discussions between the first and the last authors in which the initial codes were reviewed and refined through merging, renaming, and grouping codes to explore patterns of meaning across the dataset. Phases four and five, theme development and review and theme refinement, were carried out through a continuous iterative process with ongoing discussions among all authors, involving data, codes, and themes. The focus was on identifying clear patterns within each theme, ensuring conceptual consistency and coherence, and naming the themes [26]. In these later phases, photos were used to provide an overall interpretation of physical activity maintenance and to develop themes. For example, a reflexive discussion around a photo of a soap bubble and the story aligned with it helped capture an overall understanding of the complexity of maintaining physical activity and the environmental influence on this process. Further, there were many photos illustrating aspects of the environment. Through reflexive discussions about what these photos and stories meant for staying physically active after an mHealth intervention, we could further develop themes. In analyzing person-environment interactions, an existing behavior change theory, SCT [19], was used to inform data interpretation and further develop themes in line with reflexive thematic analysis [26,27]. For example, photos and stories about social and physical environments were useful for understanding aspects such as reinforcement and the role of social support in the first and third themes. In this way, the photos not only contextualized participants’ narratives but also contributed analytically by expanding the understanding of barriers and facilitators to maintaining physical activity after the intervention. In phase six, writing up, the final report was produced, and all authors read, discussed, and contributed to the writing through several rounds of the manuscript.

Ethical approval for the study was obtained from the Swedish Ethical Review Authority (dnr 2020‐05062 and 2022-07270-02). The research followed the Declaration of Helsinki, and careful consideration was given to ensure the integrity and well-being of the participants throughout the research process. Prior to their consent, all participants were provided with verbal and written information about the study, including their right to withdraw their participation at any time. No compensation for participation was provided. After transcription, the data were de-identified by removing identifiable information from the transcripts, and photographs were selected to ensure participants’ confidentiality.

The overall understanding of maintaining physical activity is symbolized by a photo of a soap bubble, first illustrated by a participant emphasizing the inevitable fragility of staying physically active (Figure 1): “For me this soap bubble is the a...almost impossible, a soap bubble doesn’t last forever, I mean it really doesn’t. But for how long can you hold on, for how long can you actually hold on” (female participant, 67 years old, TIA). Intervention mechanisms and personal factors contributed to the persistence of the soap film. At the same time, contextual conditions created an environment that kept the bubble afloat, even in the face of barriers in participants’ life situations. We generated 3 themes through analysis, illustrating the barriers and facilitators to maintaining physical activity after completing the i-REBOUND intervention (Table 2).

A persistent soap film served as a metaphor for gaining experience and acquiring knowledge through participation in the intervention. The continuous interaction with the physical therapist during individual counseling and regular exercise sessions over 6 months provided opportunities for habit formation. The regular engagement in exercise also enabled the participants to experience the benefits of physical activity for health and managing daily life. Along with a deeper understanding of the principles of physical exercise, the intervention components and mechanisms of change supported a stable foundation to maintain regular physical activity after the intervention.

The regular physical exercise sessions at higher intensity levels had served as a catalyst for physical activity and reinforced habit formation for continued activity after the intervention. In addition, being regularly active over time enabled the participants to experience and recognize the health benefits of physical activity. Examples of such benefits included reduced pain levels, increased energy for daily activities, and an overall feeling of improved well-being. The catalyst for change through these experiences was felt to be beneficial regardless of previous engagement in physical activity. One participant shared how the intervention was an awakening of the benefits of being regularly active: “I mean, I did move around before as well, but this has still given me a bit of…a little step forward, in some way. I think so. Uh, a wake-up call. That this is…this is good. It’s good for me to do these things” (female participant, 76 years old, stroke). In line with experiencing health benefits, the participants also became aware of the importance of increased strength for managing daily activities. This awareness was reflected in a photo of stairs (Figure 2). The participant lived in a building without an elevator, and the photo symbolized how regular physical activity was a prerequisite to sustaining the ability to climb stairs and accessing the apartment. This example illustrated how maintaining physical activity was perceived not only as beneficial for health but also as a means for sustaining important abilities for their daily life.

During the intervention, the participants developed an understanding of the importance of exercise intensity and the benefits of having an increased heart rate during physical exercise. Participants used this knowledge to adapt their physical exercise routines to meet recommendations for physical activity. Moreover, participants learned how to use daily activities as opportunities for physical activity. Walking, in particular, was an important and easily accessible activity that could be adapted, for example, by replacing slow-paced dog walks with brisk power walks without the dog. Another example of modifying an existing routine, illustrated by a photo of a swimming pool (Figure 3), was the incorporation of faster swimming intervals to enhance intensity. In addition, the introduction of tools, such as the Borg scale [35], and specific exercises provided a foundation for a personalized exercise routine and set the stage for progression. The familiarity and simplicity of the intervention’s exercises supported maintenance, as they were easy to remember and carry out, and the Borg scale was a practical tool for the participants to monitor and achieve their target intensity levels.

The participants entered the intervention with different levels of prior knowledge for their knowledge acquisition. Participants with previous exercise experience could continue to develop knowledge, and the intervention’s supervised support reinforced their understanding. In addition, the supervised sessions helped these participants develop realistic expectations for physical exercise, adapted to their age and current physical ability. One participant reflected: “This thing about setting the training at a, yeah…a level that’s just right. That made it possible to do things differently. It just felt…I mean, mentally it was different” (male participant, 70 years old, TIA). As the quote illustrates, the knowledge acquired during the intervention prompted a shift in mindset that facilitated acceptance of exercise at a more manageable level. For the participants with limited experience of physical exercise, the intervention served as an introduction to establishing exercise routines and contributed to the acquisition of knowledge on exercise and physical activity, which supported their ability to maintain physical activity after the intervention.

The participants’ context—including dynamic interactions with their social and physical environments—was crucial for keeping the physical activity bubble afloat after the intervention. Within their immediate context, they identified significant others who could ease the transition from the physical therapist’s supportive role, as well as meaningful places that offered favorable conditions to maintain physical activity.

Being part of the intervention entailed continuous social support for physical activity over 6 months. At completion, the participants missed both the guidance from the physical therapist and the social context provided by the regular exercise sessions. However, the participants identified alternative resources within their social context that facilitated the maintenance of physical activity after the intervention. One participant shared how taking part in the intervention made him realize the importance of having a companion to make it to the gym and identified a key person within his social network. He said:

In this way, having scheduled sessions with a companion fostered a sense of accountability, as someone was expecting the participant to attend. In addition, engaging in physical exercise with a companion provided a person who checked in on progress, contributed to a sense of safety, and gave feedback—similar to the support offered by the physical therapist during the intervention. Exercising with another person also brought joy and added value beyond physical activity itself. One participant described how the exercise class helped her build relationships and make friends, which facilitated her continued participation. Moreover, encouragement from significant others, such as friends, neighbors, and family members, facilitated maintaining physical activity by reinforcing the participants’ decision to stay active and strengthening their sense of doing something positive for themselves. One participant illustrated this motivational support by quoting how her children encouraged her to prioritize engagement in physical exercise:

Alongside the social context, having a designated place to engage in physical activity served as a valuable resource that facilitated maintenance following the intervention. One factor for this place was accessibility, including geographical proximity and ease of access via transportation, contributing to a lowered threshold for exercising, as illustrated by one participant: “I like it when it’s nearby and easy to get to and all that. That’s kind of my problem—when it gets complicated or takes longer than I had in mind, I lose motivation” (female participant, 67 years old, TIA). The outdoor environment was one such example. In addition, engaging in physical activity outdoors provided mental rest through its calm and quiet atmosphere—something particularly valued by participants experiencing fatigue. It could also provide a sense of familiarity; one participant described feelings of comfort in her outdoor walking routine, as she was well-acquainted with the route’s length and the time it typically took to complete. In one photo of a gravel trail (Figure 7), a participant captured how outside places could offer a meaningful nature experience and elaborated on how seasonal changes and the beautiful wildlife inspired her to take daily walks. However, using outdoor environments for daily physical activity was also associated with barriers related to seasonal and weather changes, especially rainy days or when the roads were slippery in the winter. For the participants, these barriers led to disruptions in their exercise routines—sometimes resulting in doing less than planned and, at other times, failing to adhere to the goals they had set at the end of the intervention. To overcome the weather and seasonal challenges, the participants needed to adapt their physical activity behaviors. One participant shared a photo of a kicksled (Figure 8) that exemplified this adapted behavior and how a kicksled enabled outdoor physical activity during the winter. For others, a pair of stable walking shoes or walking sticks helped maintain physical activity, through providing safety and inspiring them to go out for a walk regardless of the weather.

This study provides insights into barriers and facilitators of maintaining physical activity after completing an mHealth intervention from the perspective of individuals poststroke or TIA. The findings show how maintaining physical activity following an mHealth intervention is founded on a reciprocal relationship among the personal factors, behavior, and the environment. These findings align with behavioral change theories such as SCT [19], and the discussion will focus on aspects of personal, environmental, and behavioral factors for maintaining physical activity after an mHealth intervention.

This study found that personal factors, including positive experiences with regular physical exercise and knowledge acquired through interaction with a physical therapist, supported awareness of the health benefits of physical activity and habit formation. Consequently, these factors helped individuals incorporate physical activity into their daily routines and adjust their behaviors to follow physical activity guidelines. Similar findings were reported by Wahlich et al [36] who found that inactive mid-life and older adults made lifestyle changes in their daily lives and increased their awareness of the importance of physical activity after a pedometer-based walking intervention. Although confirming the findings of Wahlich et al [36], our study also demonstrated that participants not only acquired knowledge but also actively applied it by developing and adapting individual exercise routines after the intervention. This process was supported during the intervention by the creation of a social environment, which included support provided by the physical therapist who had extensive experience in stroke rehabilitation and played a key role by providing knowledge, reinforcement, and encouragement to maintain being physically active. After the intervention, this external support provided by the social environment was succeeded by participants’ own experiences with positive outcomes, such as reduced pain and improved well-being, which, in line with SCT [37], further reinforced them to maintain regular physical activity.

The discontinuation of support from the physical therapist was perceived as a barrier to continued engagement in physical activity after the intervention. Supervision is recognized as an important component to supporting behavior change in individuals poststroke [38] but is inherently time-limited and resource-intensive. These constraints highlights the importance of identifying and strengthening alternative support once structured supervision ends. Our findings show how participants identified resources within their social network to substitute the physical therapist’s supportive role during the intervention. These results are in line with previous findings highlighting that encouragement from significant others (eg, neighbors, friends, and family), as well as engaging in physical activity with a companion can facilitate maintaining physical activity [17,18,36]. The transition of support to the individual’s own social network suggests, consistent with SCT [39], that individuals poststroke or TIA can sustain behavior change by developing alternative sources of social support within their immediate context following an mHealth intervention. This is an important finding, highlighting the need to incorporate strategies and training into mHealth interventions that help individuals integrate physical activity into daily life, supported by their social and environmental contexts.

Participants developed and used various personal strategies, some introduced during the intervention, to facilitate maintenance of physical activity. One such strategy was the self-monitoring of physical activity, where participants tracked their activity levels. The reported preference for using step counters on their mobile phone to monitor their physical activity was consistent with findings on strategies that facilitated physical activity maintenance in adult and older populations [17,18]. Handwritten exercise logs were preferably used, while the self-monitoring tool within the app was used to a lesser extent, possibly because it did not fully meet the needs of people poststroke or TIA. Our findings align with a previous quantitative evaluation of the mHealth version of the i-REBOUND intervention, which showed a decline in engagement with the app’s self-managed tools over time [40]. This decline may reflect limited integration of key features known to enhance app engagement, such as personalization, adaptive content, and continuous feedback [41]. Although this study provides new insights into personal and environmental factors influencing maintenance—particularly those not directly related to the digital tools—further research is needed to better support long-term engagement in physical activity through mHealth solutions after stroke or TIA. Future refinements of the app may benefit from the development of more personalized content with regular updates and from the integration of automated self-monitoring of daily step counts or minutes in moderate physical activity, thereby improving its usability and relevance for this population. In SCT, self-monitoring is an important self-regulatory strategy that enables individuals to track their behavior, evaluate progress, and make necessary adjustments to achieve their goals [42]. In addition, SCT also emphasizes that maintaining physical activity requires the development of other self-regulatory capabilities, such as goal-setting and planning, as well to help individuals navigate barriers and maintain motivation [39]. These self-regulatory strategies were reflected in this study, demonstrating how participants applied the knowledge acquired from the intervention, along with their prior experience, and developed strategies for overcoming barriers. However, this study also showed that participants with limited prior knowledge of physical exercise and limited personal resources to develop strategies face greater challenges in implementing strategies for maintaining physical activity.

This study demonstrated that health-related challenges, including comorbidities, short-term illnesses, and stroke-related symptoms, were substantial barriers to maintaining physical activity after the mHealth intervention. The impact of comorbidities is an especially important consideration for health promotion interventions targeting individuals poststroke or TIA, as comorbidities are more prevalent in this group than in the general population and have a substantial negative influence on both functional outcomes and mortality [43,44]. In addition, comorbidity is strongly linked to age, with most adults older than 65 years having one or more diagnoses, and because people poststroke or TIA are often older, age-related comorbidities are frequent [45]. Furthermore, consistent with prior research [36,46,47], the participants struggled to allocate time for physical activity along with other responsibilities and engagements. These findings underscore that maintaining physical activity could be challenging for people poststroke or TIA given the outlined health-related and contextual barriers. Hence, these circumstances should be carefully considered in the design and development of mHealth interventions. Building on this study’s findings that personal strategies can help overcome health-related and contextual barriers, mHealth interventions should include components that support the implementation of such strategies in complex life situations. These components should balance the reinforcement and adaptation of existing strategies among experienced and resourceful participants, with targeted support for those who face greater challenges in the implementation.

Our results highlight the important consideration of the interaction between individuals and their environment while supporting the transition from therapeutic support to self-management in mHealth physical activity interventions for people poststroke or TIA. One potential approach is to include a structured maintenance phase in the intervention design, during which therapeutic support is gradually reduced. In this phase, participants can progressively take more responsibility for their own development and independently apply self-regulation techniques to sustain physical activity. Because behavior maintenance is a dynamic process, discontinuities are to be expected; therefore, training participants to manage setbacks is recommended [39], including strategies that foster agency and resilience to overcome health-related and contextual barriers. Our findings further indicate that support from significant others, through encouragement and companionship, is essential for maintaining physical activity. Accordingly, future mHealth interventions should be aware of the complex situation for this population and help participants identify personal strategies in daily life and engage with a supportive social environment beyond the formal intervention setting. Finally, the study findings should be interpreted within the Nordic context, as, for instance, different climates, seasonal contexts, or digital health care systems may yield distinct barriers and facilitators.

A major strength for the study’s trustworthiness was the design and methodology, which contributed to the rich dataset developed through repeated semistructured interviews at 3 months and 6 months postintervention and participant-generated photographs during the second interview. This approach facilitates going beyond surface-level descriptions of participants’ experiences by building rapport and stimulating reflection between and during sessions [23,34,48]. Further, using participant-generated photos may help reduce power imbalances between researcher and participant while also enabling the introduction of new dimensions to data generation that might not be anticipated in the interview guide, thereby contributing to a deeper understanding of the issue under study [23,34]. Furthermore, participant-generated photos and interviews at 3 months and 6 months contributed to reducing recall and response bias and to gaining both short- and long-term experience with maintaining physical activity. Additional follow-up could have provided more insight into maintaining physical activity. The design was grounded in a constructivist epistemology, with reflexivity as a central component in the generation and analysis of data [26,27]. The first author had brief contact with the participants during the feasibility trial’s data collection phase. This dual role may induce the risk of social desirability bias. In addition to the design with repeated interviews and participant-generated photos, several strategies were used to elicit authentic interview responses, such as providing assurances and probing questions [49]. These strategies for an authentic and rich dataset were developed through reflexive discussions of pilot interviews, followed by continuous reflexivity through data generation.

Limitations of the study include a lack of diversity in some of the aspects of the sample recruited from the mHealth version of the i-REBOUND intervention [21]. Diversity was obtained in terms of geography, age, educational background, and experiences with daily use of the internet and mobile apps. However, the relatively high physical activity and functional level, gender imbalance, and lack of a more complete socioeconomic background represent limitations that influence the transferability of our findings. The challenges with recruiting a diverse sample were previously addressed [21,50]. Although the participants in this study were a relatively homogeneous group, they still faced a variety of barriers related to maintaining physical activity after the mHealth intervention and needed support to adjust their physical activity to meet recommendations given their current life situations. However, our findings show important facilitators, such as knowledge acquisition and support from the social and environmental context, that can be useful for exploring conditions for more diverse populations. To facilitate the judgment of transferability to other contexts and populations, we reported detailed descriptions of the study’s methodology [27,51]. To explore how gendered and socioeconomic aspects as well as experiences of those with a more sedentary lifestyle and severe stroke influence physical activity maintenance, recruitment strategies in future studies should aim for a more diverse sample, not the least because socioeconomic disparities are shown to be related to stroke incidence and outcome [52].

This study points out the need for future mHealth interventions to better prepare people poststroke and TIA to develop and apply personalized strategies for managing barriers to maintaining physical activity. The complex life situation following stroke or TIA, involving health-related challenges and contextual barriers, can limit individuals’ ability to engage in regular physical activity. Nevertheless, participants may develop strategies, supported by knowledge and tools from mHealth interventions and the individuals’ surrounding context, to navigate these challenges. These insights highlight opportunities for future mHealth interventions to strengthen the interaction between individuals and their environment while also empowering personalized strategies for supporting behavior change and long-term physical activity maintenance.