A Mobile App (CareFit) Supporting Physical Activity for Informal Carers of People With Dementia: Mixed Methods Feasibility and Adaptation Study

Introduction

Health and social care models worldwide are facing perpetual crises. Aging populations, increasing comorbidities, and a shrinking health and care workforce create a “perfect storm” of continuous system decline [1]. Caring roles are increasingly taken on by the community, often by informal (family) carers [2]. In the United Kingdom, 10 million informal carers save the medical system £184 billion (US $249 billion) each year; the cost of a second National Health Service (NHS) [3]. NHS data reveal that carer-reported quality of life is decreasing year upon year, and recent national-level surveys in the United Kingdom (engaging >10,000 carers) indicate that 82% of carers see challenging impacts of delivering care on their physical and mental health [4-6]. There is a compelling need to recognize this paradigm shift and build much greater resilience in our care system and networks. This includes asking difficult implementation questions as to how scalable solutions for carers can be co-designed, implemented, and maintained.

Across a growing array of technology [7,8], the development of digital supports for carers of people with dementia continues to be of significant societal interest [9-11]. UK data suggest that 1 in 2 people will be affected by dementia in their lifetime—either by the condition or by caring for someone with dementia (944,000 people currently live with dementia in the United Kingdom) [12]. Understandably, dementia has long been a topic of interest for developing digital interventions, including the pioneering iSupport platform developed by the World Health Organization that has facilitated evidence-based theory such as personhood and cognitive reframing [13,14] becoming accessible to a global audience through online training and support. While iSupport delivers key material of critical importance (eg, introducing dementia, dementia-related behaviors, and carer health and well-being), there remains scope to increase the capacity of the platform to provide personalized advice to support the regular physical activity of carers. More broadly, physical activity among carers is a surprisingly underdeveloped area of current science. Multiple reviews have highlighted that carers are largely overlooked when it comes to physical activity innovations (of both dyadic and independent carer use), with systematic reviews identifying between 11 and 14 interventions to date aimed at carers, including a strong emphasis on face-to-face and telephone delivery [15,16]. Given this gap in knowledge, a team of multidisciplinary researchers at the University of Strathclyde co-designed and evaluated (with carers and health professionals) a prototype digital tool, CareFit, aimed at increasing physical activity (and reducing sedentary behavior) in carers [9]. Awareness of physical activity guidelines (eg, from the World Health Organization and the UK government [17]) can be poor among vulnerable groups [10], including carers [18]. Therefore, the platform was built with an emphasis on “actionable information,” particularly for those starting out in physical activity, and our original work was conducted during the COVID-19 pandemic lockdowns. The intervention design was largely influenced by the transtheoretical model of behavior change [11,19], is self-led by informal carers at their own pace, and does not require specialist equipment. The initial (3-week) co-design and pilot study across Scotland [18,20] successfully demonstrated the initial acceptability, feasibility, and usability of the approach, which was open to use across any carer group. Remaining gaps in knowledge include a more in-depth understanding of the practicalities of delivering CareFit within a single postdiagnostic pathway across recruitment, adherence, and outcomes. This includes critical questions regarding how to personalize the intervention, how to extend the platform beyond 3 weeks of use, and the identification of key barriers and enablers to facilitate more widespread and scalable delivery. We also sought to simplify the use of CareFit from previous studies and consider more human and community elements—particularly given key challenges regarding the digital divide [20] for informal carers. Thus, we set out to co-design, adapt, and explore the feasibility of a novel cross-platform approach to support physical activity in carers of people with dementia.

Methods

Objectives

Our aim was to co-design, adapt, and personalize a cross-platform digital health app (CareFit) designed to support regular physical activity in carers of people with dementia for a period of 8 weeks and evaluate the potential for implementation.

Our objectives were to (1) expand an initial 3-week intervention to an 8-week intervention to support maintenance of physical activity; (2) develop an understanding of recruitment pathways and explore barriers to and enablers of recruitment for a future definitive trial, including recruiting from more deprived socioeconomic groups; (3) improve the understanding of use adherence and develop an understanding of the most reliable methods to regularly measure physical activity and sedentary behavior among dementia carers; (4) explore how CareFit could provide added value for dementia carers with other existing digital interventions (such as iSupport) through qualitative interviews or focus groups with key stakeholders; and (5) explore unexpected benefits, including whether people with dementia also see the benefits of CareFit.

Study Design

Overview

The CareFit for dementia carers study was a concurrent mixed methods [21] evaluation of a novel motivational mobile app to support home-based regular physical activity for informal dementia carers. The study methodology has been published previously [22] and forms a stand-alone element of an existing randomized controlled trial of iSupport [23]. The mixed methods evaluation was informed by both the Medical Research Council (MRC) complex intervention framework [24-26] and the reach, effectiveness, adoption, implementation, and maintenance (RE-AIM) framework [27].

For initial co-design to achieve both adaptation and expansion, we used 3 development “sprints” where different elements of the app were sequentially developed with our local development partner, Add Jam [28]. Feedback was sought from a range of stakeholders such as carers and professionals as well as researchers and developers to create priority areas for app development. We were also able to build on priorities from previous work on CareFit and the feedback obtained [24]. Our subsequent feasibility study involved both local and national networks in Scotland as well as the use of Join Dementia Research (JDR) to understand the feasibility of recruiting for a larger trial. We aimed to recruit 50 carers of people with dementia for our feasibility study. This number was arrived at based on both related statistical guidance for pilot studies and anticipated high dropout [14,29].

Feasibility Study Inclusion Criteria for Carers

The inclusion criteria for carers were as follows: (1) adults (aged ≥18 years) living in Scotland who self-identified as informal carers (eg, partners, children, or friends) of a person with dementia (self-reported), (2) contemplating or preparing to practice physical activity, (3) ability to carry out simple exercises such as arm raises or stretching, (3) ability to read and write in English, (4) access to a smartphone (Android or Apple) as well as access to the internet, and (5) normal or corrected-to-normal eyesight.

Feasibility Study Exclusion Criteria for Carers

The exclusion criteria for carers were as follows: (1) anyone advised by a clinician not to undertake physical activity or make any change in their present level of exercise, (2) already regularly exercising to a significant level outside the home (eg, running or cycling), (3) residence outside Scotland at the time the study was conducted, and (4) current participation in a related iSupport study.

Feasibility Study Inclusion Criteria for Professionals

The inclusion criteria for professionals were as follows: (1) adults (aged ≥18 years) living in Scotland and working as health and social care professionals and (2) willingness to engage with the study (eg, share information about CareFit with carers of people with dementia) through their professional roles for a period of at least 3 months (to ensure timely follow-up).

Feasibility Study Exclusion Criteria for Professionals

The exclusion criteria for professionals were work exclusively within the NHS.

Ethical Considerations

Ethics approval for feasibility work was obtained through the Bangor University School of Medical and Health Sciences Academic Ethics Committee (approval 2021-16915). The initial co-design and adaptation work was supported through approval 2022 from the University of Strathclyde, Department of Computer and Information Sciences. CareFit was coded using React Native (Meta Platforms) so that we could facilitate use across both Android and Apple mobile phones and tablet devices. As the study was a “closed” research project (ie, not available freely to the public), onboarding of participants involved a number of steps regarding contacting the research team, confirming that the participants met the study inclusion criteria, assigning participants a study ID, and then sending them a bespoke link for the corresponding Apple and Google stores. This process was thoroughly tested before the implementation study began. The protocol publication provides further information [22]. Informed consent was obtained from participants in the studies after reading a participant information sheet and given an opportunity to ask questions. Participants were given a small thank you token for taking part in a study. All sessions will be conducted in a way to ensure privacy and confidentiality.

Research Team and Reflexivity

The research team conducting qualitative interviews comprised 3 (male) research assistants (WH, RM, and RI), who have expertise undertaking qualitative work within digital health research, and the project lead (male; KJE), who has >10 years of postdoctoral experience and a specialist interest in carer populations. The interviews and focus groups were facilitated by all 4 team members (WH, RM, RI, and KJE)

Data Collection, Management, and App Download Process

Study advertisements (in both paper and digital formats) were shared with study partners in parallel with the use of the JDR platform (Figure 1). Beyond monitoring study interest numbers, our key areas of data collection included (1) online surveys, (2) “in-app” data collection, and (3) interviews and focus groups. For online surveys, information was shared using the online platform Qualtrics (Qualtrics International Inc) [30]. Carers who consented were directed to a bespoke hyperlink within the Google Play Store and Apple App Store where an individual pseudoanonymized ID was entered after download. For in-app data collection, data included use and time stamps for users across the app tabs, such as Activity, Planner, Resources, and Sharing.

‎

Qualitative Data Collection

For an overview of the qualitative approach, Multimedia Appendix 1 provides the full COREQ (Consolidated Criteria for Reporting Qualitative Research) checklist [31]. The research assistants (WH, RM, and RI) had previous background experience in qualitative work within digital health research across a wide range of chronic conditions (eg, cancer, diabetes, and kidney failure). All staff members had conducted qualitative studies previously, including at the master’s in research (WH and RI), PhD (RM), and postdoctoral (KJE) levels, where training was supported by fellow staff members and university training courses. The researchers had no professional or ongoing relationship with informal carer participants, and the latter were recruited through convenience sampling—however, a small number of the professionals interviewed at the conclusion of the work were identified through university roles (n=2). All participants were informed that researchers were interested in exploring the use of digital approaches to support health and wellness for carers specifically through physical activity. We repeatedly shared with participants that there were no “right” answers and that relationships or care would not be affected in any way by the responses given. We asked all participants for basic demographic information so that we could accurately describe our study samples.

Interviews and focus groups were conducted online, and the format was internally pilot-tested before use. There were no nonparticipants present in any of the interviews, and no repeat interviews were conducted. The interviews lasted approximately 30 to 45 minutes, and we did not capture field notes for the focus groups. Audio recordings were checked against transcripts to ensure accuracy, but the transcripts were not sent back to participants for checking. Each researcher coded their own individual interviews, which were then cross-checked and discussed with another member of the team. Although we approached individuals who left the study for feedback as to why they dropped out, we did not manage to obtain such feedback.

Adaptation of CareFit

We successfully expanded and adapted CareFit for 8 weeks of use within the context of dementia care (Table 1 and Figure 2). This included developing features highlighted through previous real-world use of a first-edition app [18,20], including aspects such as a redesign or simplification of the graphical user interface. We also redesigned the physical activity planner (placed on the Home tab) so that users could both add and review activities using the same screen and add their own activities. During our 3 development sprints, we presented progress across a range of charity partners (including Alzheimer Scotland) and also invited stakeholder partners to try out elements of the app by downloading prototypes in development. Co-design outputs (Table 1) placed a strong emphasis on carer feedback from 16 carers who had previously used a version of the app for 3 weeks and told us priorities for improvement. This feedback was integrated with charity partner feedback and technical feasibility feedback to adapt the requirements accordingly. Simplification ultimately led to 4 tabs: Home, Learn, Community, and More. Following recommendations for simplification, we established a uniform image set across all activity videos and color coded these according to whether the activity was cardiovascular (red), sedentary (gold), or related to muscle and balance (blue), reflecting UK physical activity guidelines. Additional elements included short video clip animations summarizing key information across the Learn section; the development of fictional personas that described a variety of dementia caring scenarios across different genders, geographical settings, and ages; and links to local resources within the Community tab.

‎

Evaluation Outcomes

Our mixed methods evaluation was designed to improve our foundational understanding of CareFit supported by validated instruments wherever possible. This included measures to understand usability (System Usability Scale [SUS] [32], which is a widely used 10-item Likert scale to provide a global view of subjective assessments of usability), improvements in quality of life and economic impact (EQ-5D-5L [33], which is a 5-point scale covering domains of mobility, self-care, usual activities, pain or discomfort, and anxiety or depression), and changes in physical activity or sedentary behavior (International Physical Activity Questionnaire–Short Form [IPAQ-SF] [34]). The IPAQ-SF has been developed and tested for use in adults (aged 15-69 years) and asks individuals to report the number of days and time spent performing moderate and vigorous physical activity, walking, and sitting. We sought to deepen our contextual understanding of the population by capturing information on demographics (eg, that could be used as covariates in later analyses) through baseline and follow-up questions on age group, gender, and number of years of caring and hours of caring per week. Specific questions for informal carers also included information to allow for the calculation of BMI, a single question on sedentary behavior and muscular activity, and familiarity with and awareness of physical activity. Professional and carer interviews and surveys were structured around the RE-AIM and MRC complex intervention frameworks (Multimedia Appendices 2 and 3). Other data collected included in-app data, which primarily supported our understanding of intervention adherence (and dropout).

Analysis

Our use of RE-AIM focused on the planning and conduct of data gathering as opposed to the evaluation phases (see the Limitations section). To better understand uncertainty regarding key elements of the MRC complex intervention framework, we present key results structured around recruitment, intervention design, and suitability of outcomes. Our approach to mixed methods [21] was to analyze qualitative and quantitative results individually in parallel and combine them to draw conclusions. The interviews were recorded using an encrypted dictaphone and university-approved video software.

Thereafter, the thematic analysis process by Braun and Clarke [35] was followed using NVivo (version 12; Lumivero) and was guided by interpretivism and the individual perspectives that informal carers and professionals can provide. We chose thematic analysis due to particular strengths regarding exploring perceptions, attitudes, and motivations and finding deeper meanings as opposed to the more rigid framework of content analysis. Due to the diverse nature of the population interviewed, we did not anticipate reaching data saturation but, instead, invited as many professionals and informal carers to come forward as possible (through convenience sampling) within the time limits of the work. We did not conduct a formal coding tree but discussed notes and reflections together as a group. Deductive approaches were supported primarily through the use of the complex intervention framework (namely, feasibility of recruitment, intervention design, and outcome suitability). The minor subthemes presented were derived from the data throughout, and examples of quotes are provided consistently.

For quantitative outcomes, basic statistics related to feasibility were gathered (eg, description of key recruitment numbers and adherence to CareFit from our in-app data, which included summaries across the Activity, Planner, Resources, and Sharing elements with time stamps of the time and date of use). Indicators of usability were underpinned by the SUS total scores, and on occasion, individual questions were used to support the interpretation of qualitative data. For all outcomes, differences in measures between baseline and follow-up were analyzed primarily regarding metrics that supported our understanding of feasibility and suitability of outcomes, such as completeness of the information.

Results

Description of Feasibility Study Populations

Our baseline demographics data describe 41 participants recruited (Tables 2 and 3). Most carer participants (39/41, 95%) were female, and all participants (41/41, 100%) self-identified as being White. A total of 63% (26/41) were educated to a degree level, and age groups varied from 25 to 34 years up to 75 to 84 years. Almost half (18/41, 44%) of carer participants had been caring between 3 and 10 years. Almost a third (13/41, 32%) reported caring for ≥8 hours per day, and 59% (19/32) of the council areas were represented. At baseline, the general adiposity (measured through BMI) varied from healthy (13/41, 32%) to morbidly obese (4/41, 10%). Activity levels as per the IPAQ-SF scoring were relatively equal across the vigorous (12/41, 29%), moderate (15/41, 37%), and low (14/41, 34%) groups. Muscular activity guidelines (≥2 days) were met by 39% (16/41) of the participants at baseline.

A total of 27 professionals participated in the interviews or focus groups, and 25 (93%) completed the detailed follow-up survey on the feasibility and implementation potential of CareFit. This population included pharmacy staff, those working in public sector roles across health and social care, charities and the third sector (at the national and local levels), public health roles, and digital health professionals, as well as those working in political roles. We do not detail each position individually to retain the anonymity of the participants (Table 4).

Feasibility of Recruitment

Overview

We contacted 659 volunteers from JDR. In total, 88 carers of people with dementia (including n=34, 39% from JDR) showed an initial interest in the study, and 41 (47%) of them progressed to completing the baseline survey. A total of 41% (17/41) of our total study population was recruited from JDR. For the 24 carers who were recruited through direct advertisements, recruitment took place most notably through Carers Scotland, social media, and Alzheimer Scotland or Brain Health Scotland. The most successful nonregistry recruitment approaches were emails from the research team (23/40, 58%) and social media posts (14/40, 35%). Table 2 provides further information.

Qualitative data further improved our understanding through the identification of the 5 key subthemes outlined in the following sections. Table 5 provides an overview of the subthemes and further quotes.

Lack of Carer Self-Identification and Related Terminologies

A key challenge for recruitment was that people who care for others do not always identify with terminologies such as carer or caregiver. Such challenges are extended further across more demographic-specific barriers to research and access to health care related to gender or cultural background, personal viewpoints, or cultural beliefs:

The trouble, I suppose, is that you know, with carers, a lot of carers don't badge themselves in that way.
[Professional participant 5]

Digital Readiness of Informal Carers of People With Dementia

Stakeholders typically assumed carers of people with dementia to be older, less familiar with technology, and difficult to engage due to caring responsibilities.

The level of responsibility placed upon carers of people with dementia was highlighted by a number of professional stakeholders. The diversity in technology confidence and support requirements was highlighted across stakeholder interviews:

One of my questions might be is from the Alzheimer’s carers’ audience, they might be the ones that are in most need, but they might also be the hardest ones to actually get.
[Professional participant 8]

Catering to Different Stages of the Caring Journey

Targeting earlier provision in the journey of a carer was a common suggestion, ensuring that the intervention becomes part of postdiagnostic support. This could include linking with general practices, health centers, and carer centers. The consensus was that, as dementia progresses, caring needs placed on the carer will become too demanding to allow CareFit to be as effective or easy to implement. Linking with other services and organizations would be effective at the local level, but wider integration could be achieved if CareFit was able to align with national and regional policies, strategies, and agendas.

Increasing Breadth and Depth of Advertising to Identify Informal Carers

Suggestions from stakeholders regarding advertising varied considerably and included advertising approaches both in person and at distance, such as the use of local radio advertisements, social media, visibility on both the Apple App Store and Google Play Store, telephone contact, leaflets, and posters as well as more unstructured approaches such as word of mouth. In terms of the content of such materials, stakeholders mentioned that the benefits of using CareFit could be important to highlight upfront.

Coordination With Key Stakeholders, Including Using Established “Watering Holes”

Many stakeholders highlighted the need to coordinate at a local level with other services that carers already frequent, including “less formal” networks that may involve third sector presence and established local services. More broadly, stakeholders highlighted that face-to-face roles remain important for app uptake and use, where roles such allied health professionals or the pharmacy sector may complement future implementation. To a lesser extent, stakeholders also had some suggestions for coordination around national-level and local-level initiatives.

Feasibility (Safety and Usability) of the Intervention

We first examined the feasibility of the intervention through the lens of safety—there were no adverse events reported by any of the participants involved in the study. In terms of usability, SUS responses were completed by 44% (18/41) of the participants. The average score was “acceptable” [36]. On the basis of the responses to the individual questions, participants were primarily positive toward the CareFit app (eg, 8/18, 44% strongly agreed that they would like to use CareFit frequently [question 1], and 15/18, 83% disagreed or strongly disagreed that CareFit required learning a lot of things to use it [question 10]).

Adherence and Practical Considerations for Use: Informal Carer Data

Overview

A total of 41 participants initially joined the study, 38 (93%) downloaded the app, and 33 (80%) used it throughout the 8 weeks. In total, 51% (21/41) of the participants submitted the follow-up surveys (see the Methods section for further details). We explored use through 8 weeks of content management system data. Our overall “tag” counts (from 2891 instances) identified that the most prominently used parts of CareFit were the Learn tab (1296/2891, 44.8%; 1296 events), followed by the Activities (966/2891, 33.4%) and Planner (230/2891, 8%) tabs. Other available tabs were used to a lesser extent, such as Own activity (163/2891, 5.6%), Sharing (159/2891, 5.5%), and Resources (47/2891, 1.6%).

Events and associated time stamps (Figure 3) identified a large initial engagement with CareFit, particularly with the Learn tab (eg, week 1), as well as a subset of users who established steady engagement with the Activity section throughout the 8-week study. Small peaks of app use could be observed at the 3- and 5-week points. This timing coincided with 2 keep in touch emails sent directly to users on the platform to offer technical and then more general study support. Figure 4 shows use across different times of the day, where we observed higher use during preoffice hours, a reduction in use in the middle of the day, and late-afternoon to late-evening use.

‎

‎

We further explored adherence and practical considerations through themes generated from carer data (21 questionnaire responses and 8 qualitative interviews). We summarize these data through 5 key subthemes in the following sections (Table 6).

Caring Routines

Regardless of the physical activity level achieved during the study period, most carers highlighted that accessibility to physical activity was associated with the demands of the caring role and being able to partake in activities “when you want.” Initiating physical activity around busy routines was something mentioned as a specific challenge to carers, where changes were supported through an appreciation of positive impacts, including those in the short term:

No, I enjoyed it. I didn’t use it probably as much as I should have, but that’s just because of things going on in my life.
[Carer participant 1]

Motivation for Physical Activity

Motivation for physical activity was an important point raised by many carers. While many carers felt that they knew why practicing physical activity would be important, both the impact and uncertainty of the caring role limited the opportunity for carers to practice regular physical activity. Other carers also commented on barriers to motivation for physical activity, such as feeling “exhausted” or environmental challenges such as poor weather. A number of carers mentioned that the app had helped change their perceptions of physical activity. Study adherence and practical considerations were explored through interviews and follow-up surveys:

I really enjoyed [using CareFit]. I thought it was easy to use and it fitted round about my caring duties. I found it very motivational at times and it made me think about my health. My physical health which I’ve ignored.
[Carer participant 4]

Accessibility and Usability of CareFit

Accessibility and usability of digital devices was a prominent topic in carer feedback. For example, first and early use of CareFit involved a number of steps to follow directed by the research team. Among 18 carer replies, installation was described as “easy” or “fine” by 15 (83%) participants, with 3 (17%) noting some issues or need for technical support:

[You] made it really simple. I work in a in a fairly similar role so I understand what the need for satisfying ethics and making sure that everything’s done so, no, I thought you did it very well actually.
[Carer participant 2]

Regarding longer-term use of CareFit, systemic challenges regarding technology (eg, mobile signal) were raised by participants, as well as more specific challenges regarding the design of CareFit.

Over the course of 8 weeks of use, a number of carers highlighted specific technical challenges regarding the use of CareFit. One challenge highlighted by a number of participants was the size of the physical activity videos on a phone screen, particularly when viewed at a distance. This included the need to maximize video size by putting the phone in a landscape orientation:

...the size of the screen that they offer for some of them. You know, and put your phone this way, put your phone that way. And it’s like when my phones way over there and I can’t quite see it.
[Carer participant 7]

In the earliest stages of the study, there were also specific technical issues with the activity planning and recording section of the app that a number of participants commented on. We also received more broad feedback on accessibility, including potential future challenges such as those regarding costs, where the app being free at the point of use would be particularly attractive.

Content Engagement

When asked which feature or part of the CareFit app was most useful, the survey responses received highlighted the activity videos (13/18, 72%). Other responses mentioned the activity planner (3/18, 17%), Community tab (1/18, 6%), and Learn tab (1/18, 6%). Notifications were suggested to hold utility in reminding participants to use the app, particularly in the earliest stages of use, but were also associated with pressure. The activity planning and recording was highlighted as a key strength by a number of participants. There was a range of comments regarding the Learn tab and associated content. Some carers found such information beneficial, but others pointed out that they felt that they already knew enough information to start on a physical activity plan before first use.

Suggestions of Improvements to CareFit for Future Use

A number of participants suggested ways to improve CareFit going forward. As mentioned previously, this included technical improvements suggested for the activity planner. Feedback included a focus on integrating additional devices and additional physical activities, increasing community elements, and improving aspects regarding the reminder notifications. For example, the use of a laptop would allow for a better screen size for participants to use.

Adherence and Practical Considerations for Use: Professional Data

Overview

Study adherence and practical considerations were explored through interviews and follow-up surveys with professionals who worked with dementia carers. The 5 key subthemes identified are listed in the following sections (Table 7).

Addressing Different Levels of the Digital Divide

Stakeholders highlighted that the initial adoption of CareFit requires overcoming numerous barriers, including different levels of the digital divide. For example, dependency on having suitable technology and means to access the app can empower some and limit others. Connecting through Wi-Fi and mobile data or potentially needing to pay for a regular subscription to access CareFit may not be viable if carers’ finances are limited. Such barriers also extend to human and wider cost elements:

...it gets down to you actually have to practically show people how to use things. So you know apps being solutions on their own I think is never the answer. You need to have human interaction to show people the benefits of these things...
[Professional participant 6]

Digital literacy and skill were noted as potential barriers, especially as carers are typically assumed to be older and any new technology requires time to build familiarity and skill, among existing burdens. The data privacy and quality of the app would need to be guaranteed as well. There was also the recognition that some carers will simply not wish to engage with digital interventions and prefer traditional offline alternatives. At the same time, stakeholders felt that the prevalence of technology and recent experience of providing services online throughout the COVID-19 pandemic demonstrated that these concerns can be overcome in this population. Creating an offline version of CareFit reassured some concerns.

Digital divide concepts also extended to existing routines and practices. For example, carers may use other digital resources whether they choose to or out of necessity, but others may see them as a further burden on them (ie, another account and another password to remember). Digital literacy and language barriers may also prevent carers from accessing the support that CareFit can provide, and the app would require an accessible and usable interface to be effective (eg, minimal use of text, with interactivity and images being more important). Given the limited time and opportunity that carers have for physical activity, streamlining the process of using the app to complete intended tasks should be prioritized (eg, minimal steps to starting an activity).

Timing and Target Group for Intervention

Stakeholders were cognizant that creating new habits in carers requires careful consideration of the transient nature of caring and that presentation of CareFit at an optimal upstream timing was critical:

I see the caring journey in stages cause we try to adapt our thinking not to what we do, but what they do. So many organisations really get involved with carers when the caring is really critical and impactful, at which point are very unlikely to have time to think about the app, so you would want to go to the downstream, the green where the early intervention and getting things in there.
[Professional participant 7]

A critical barrier to achieving regular physical activity remained regarding habit and the ability to change the societal norm that carers are the last people who will look after themselves given their existing pressures and commitments.

The expectation placed upon carers that they should be exercising and using interventions such as CareFit has the potential for harm or burden, with stakeholders suggesting that carers could experience a sense of guilt if they could not effectively do so among other caring responsibilities. Guilt may also arise from the idea of taking time for themselves rather than focusing on caring despite the possible benefits that self-care can offer for both the carer and the person cared for. Carers are limited in available time and resources, especially as the condition progresses. They may also simply value other ways to use their downtime instead of exercising or practicing physical activity. Culture and language differences also play an important role in physical activity engagement.

Facilitators of habit formation included both designing an app that accommodates rapidly changing situations due to events such as hospital admissions; illness of the carer or the person cared for; and carers’ anticipatory grief, especially in later stages of dementia. All carers may face some limitations regarding opportunities for physical activity, and between individuals, the rate of behavior change, personal goals and end points, and activity levels can vary, requiring an adaptive approach to enable habit formation. Equally, the caring role can also change when the carer transitions from their role (eg, the person cared for passes on or the responsibility of caring is taken on by someone else) or if the carer reaches a stage where they no longer depend on CareFit to continue their new physical activity behaviors.

Integration With Other Services

Stakeholders stressed that CareFit should not be delivered in isolation but, instead, should leverage value and visibility from a wide variety of existing services at local, regional, and national levels to integrate into a suite of existing tools. However, similarly to many novel interventions and services, there remains a risk that the promotion and use of CareFit will become another thing to do that may not be practical given current workloads and roles:

If you’re able to bake that into what they’re already doing. Then you’re kind of winning on it.
[Professional participant 16]

Suggestions for how such an intervention could be both introduced and maintained included a number of professional roles and services. For example, this could include support from professional staff on house visits where informal carers would be present.

Feedback on integration included established digital health resources and tools from both NHS providers and dementia or caring charities. An example was local council physical activity websites or health and social care mobile apps (eg, the Jointly app developed by Carers UK).

Suggestions regarding integration also extended to wider digital commercial sectors and “consumer” marketplaces. Given that the market for physical activity and digital is already busy, connections to platforms such as Apple Watch and Apple Fitness+, Fitbit, and Strava can make interventions such as CareFit more attractive to users, especially those already using these different physical activity tools.

The Need to Embrace Human Elements and Increase Social Connections

Alongside suggestions of integrating the app into services and staff roles was also a broader need to embrace human elements across the trajectory of care. The persistence of the app across the carers’ local communities and environments was a strategy proposed to secure engagement through demonstrations, testimonials from “champions,” and presence at regular activities and events within the community (eg, Digi-PALS library technological support, coffee mornings, or social events). At the onboarding level, this included critical conversations that professional stakeholders such as pharmacists have with carers, including “chatting them through what care fit is and how it could benefit them.” Motivators for carers to adopt CareFit included carers championing their own use, stories with case studies, and testimonials highlighting the benefits and minimal demand on time and energy required to use CareFit. The potential value of CareFit could also be expanded to directly reduce isolation, possibly through participation with other carers in group activities and social features or even as a topic of discussion within the community. Linking to existing platforms, including those outside of the health and social care settings (eg, social media, Alexa, and Strava) could also encourage greater adoption by carers:

So you know and maybe monitor your CareFit progress or how you’re using it, how’re you going and you know make it more of a kind of social element to it. So I suppose that’s a lot of people, you know, like that team type thing with their friends and stuff like that. So any opportunity to maybe kind of you know make it with a kind of social thing would help at all.
[Professional participant 3]

Facilitators to Support Implementation and Longer-Term Use

Facilitators to support implementation and longer-term use included suggestions at both the professional organization level and the individual informal carer level. At the organizational level, addressing the financial and structural models was raised by a number of stakeholders. This included generating knowledge of the organizational infrastructure to establish which key decision makers could initiate use of CareFit even on a small scale. The cost, resources, and effort to maintain such an intervention were also raised by stakeholders, more so if they aimed to provide technology and means to access the intervention as well as training and support.

Organizations may also be hesitant to implement new technology, with the development, provision, and maintenance of digital interventions imposing financial costs and more demand on already burdened staff who, similarly to carers, may be opposed to the introduction of new technologies and approaches. A strong evidence base and guaranteed uptake with users is required to assure organizations that there is minimal risk in implementing the proposed technology and that existing inequalities will be addressed rather than exacerbated (eg, having to provide access and connectivity to use an intervention for those who may not have the means).

Professional stakeholders largely appreciated that longer-term delivery of CareFit would require financial sustainability, although there was not a consensus on how this could be achieved:

As decision-makers and commissioners as well as policy-makers, if an intervention had clear benefit it might be able to benefit from some kind of national funding. But more likely that we would expect local providers to fund from existing budgets.
[Professional participant 11]

I think the payor is most likely to be the public sector but ideally CareFit is being delivered as a business solution or service to that market.
[Professional participant 8]

At the individual carer level, several suggestions were made to support the use of CareFit in the longer term, including beyond the initial 8-week study outlined in this paper. First, CareFit must be accessible and available to carers, whether this is through incorporating it into a suite of existing tools and resources or advertising through leaflets and posters with QR codes. For specific devices, CareFit should be available through app stores and use native functionality and services to ensure that the app is passive and requires minimal effort to start using it. Updates and new content would also be required to keep interest in the app but should not compromise familiarity or fundamentally change the CareFit app, which may discourage users instead. This would ideally make carers’ feedback result in effective changes and improvements and keep information and content up-to-date and relevant.

Prompts, reminders, and notifications from the app itself could be important tools to ensure consistent engagement throughout long periods but should be designed to be motivational and encourage the user to want to engage in physical activity rather than just reminding them of a need to complete another task. Gamification and rewards such as challenges and goals where a defined target needs to be reached can similarly encourage engagement. Rewarding continued engagement, such as streaks (eg, completing an activity each day or week), can also be essential to habit forming but can be dismaying when disengagement, which may not be possible for this user group to control or predict, is punished. Competing with other users, such as leaderboards or group progression toward a shared goal, were also suggested as engaging features suitable for some users.

Social connectivity was described in other features besides competing, such as social feeds or forums to share successes and receive support from peers on the app or using physical offline groups, activities, or classes to encourage use and check in on how CareFit is working for carers. These ideas were particularly supported as the caring role can be isolating and provide limited opportunities to connect with others. There were also suggestions to integrate physical activities that could be completed by both the carer and the person living with dementia (eg, chair-based exercises or joint walks), removing the need to find separate time and space.

Feasibility and Suitability of Outcomes

Overview

We explored the feasibility and suitability of outcomes across outcome measure response data completeness and the subjective utility of measures through professional stakeholder interview feedback (including unintended consequences; Table 8). Outcome measures were broadly well completed at both baseline and follow-up (Table 9). Some scales such as the IPAQ-SF showed completeness for some but not all questions; for example, measures of walking activity and sedentary behavior had a >90% completion rate. However, IPAQ-SF vigorous and moderate activity measures were less well completed, with response rates of 81% (17/21) and 76% (16/21), respectively, at follow-up. It was not clear whether the omitted values were because individuals did not complete activities within a specific intensity category or because they missed the question altogether. We included a question on breaking up sedentary behavior as well as a question on muscle and balance (Multimedia Appendix 4) as these are included components of the current physical activity guidelines and were integrated into the CareFit app. Both our question on breaks from sedentary time and muscular activity (days) question were relatively well completed, with response rates of >75% at baseline and follow-up.

Professional stakeholders highlighted both short-term and utility metrics of interest, including those based on measuring usability, including from professional perspectives; direct improvements in physical activity levels and sedentary behavior; broader (secondary) impacts and unintended consequences; and inclusiveness of marginalized groups and those facing inequalities.

Measuring Usability, Including From Professional Perspectives

Initial assurances for uptake included usability, ease of use, and acceptability of CareFit. For example, early evidence for use could be obtained by facilitating continued feedback from users and testing with disability organizations or those that advocate for the needs of people with disabilities. Stakeholders were largely reassured of the quality of the project through the incorporation of co-design from the outset. They suggested that use cases and real experiences of users would be useful demonstrations of effectiveness and ease of use.

Measuring Broader (Secondary) Impacts, Including Unintended Consequences

In terms of measuring changes over time, stakeholders were interested in seeing increases in initial physical activity levels, particularly continued or maintained physical activity. While physical activity increases are recognizably positive, providing support for users to understand the outputs and results of CareFit is also necessary to assist them in sustaining such behavior in the long term. Delivering the right type and level of support to an individual is also important. This may involve linking with other services, such as more intensive referral programs that run for several months before users reach a stage in which tools such as CareFit become viable for more independent exercise and activity.

Other broader outcomes described included quality of life, well-being, number of crisis points (ie, unplanned admission of the carer or the person cared for into hospital settings), hopelessness, frailty, and mortality improvement measures. These measures may also provide helpful insights apart from measuring just physical activity and help provide a better representation of an individual’s progress. The topic of person-centered support and informed goals was also mentioned when discussing effectiveness, giving control of the goals and progress to the individual carer rather than applying formative assessments to the population as a whole. Stakeholders often discussed how individual carers presented with needs for different types of support (eg, health and well-being, legal, financial, and emotional) alongside their own goals. Addressing this with individual carers was often the approach of many stakeholders, with personal goals still linking with the goals of the wider service or organization. Empowerment of carers and the building of confidence and skills for effective self-management were encouraged:

So move to personal outcomes, a formative assessment, have people—what you’re asking people to invest in “what are you gonna do, not what we’re going to do for you.” What are you going to do for yourself? And then benchmark that, three months ago, or whatever time scale, a week ago. You said you would do this. How have you done this week?...But empower and enable people to be in control of that rather than do it to them.
[Professional participant 7]

From professional feedback, common unexpected consequences of CareFit that were discussed included increased carer resilience, more effective self-management, and more confident and able carers.

The subject of unintended consequences was additionally explored through carer feedback. While most carer data did not suggest any unexpected benefits, feedback included the person with dementia taking part in physical activity, completing physical activities but not noting the activity within the app, and becoming more competitive with themselves as carers.

Inclusiveness of Vulnerable Groups and Those Facing Inequalities

While promoting self-management and resilience may benefit the system supporting carers, concerns were raised that marginalized and vulnerable groups would continue to struggle even if CareFit were to be successfully implemented. Directly addressing a range of public health behaviors in carers through the use of CareFit would be valuable to sectors such as public health and for gaining support but would require long-term use and longitudinal data collection with thoroughly planned methods and measures. While dementia is a long-term condition and there is potential to support carers over these periods, unpredictable routines can make continued regular use hard to guarantee and enforced timescales less feasible.

Discussion

Principal Findings

We successfully co-designed, expanded, personalized, and evaluated the implementation potential of a mobile health app to support physical activity in carers of people with dementia. Throughout, our approach was to learn about the real-world practicalities regarding implementation guided by both the MRC complex intervention framework [24-26] and the RE-AIM framework [27]. Our co-design outputs built significantly on existing research through establishing a greater depth and improving the functionalities of CareFit. By building a bespoke content management system, we established a future capacity within the platform to capture real-time knowledge of carer well-being and deliver key messages to carers in local community settings, including at scale. Self-identification regarding being a carer appears critical for uptake—early introduction of CareFit, including through face-to-face interactions, appeared extremely useful. Intervention uncertainty findings highlight positive feedback on the graphical user interface; however, dropout of carers at 8 weeks was notably high at 49% (20/41; 21/41, 51% of the participants completed the study). Finally, while completeness of outcome measures was relatively good, there remained gaps regarding key validated measures such as the IPAQ-SF. We included 2 novel measures of sedentary breaks and muscle and balance activities, and these were largely well completed. These are important components of the current physical activity guidelines that remain particularly easy to overlook.

Limitations

Despite a number of achievements, several limitations remain. Due to time constraints, we did not apply for UK Health Research Authority approval to recruit participants in receipt of NHS services. This additional recruitment channel may have increased reach and helped us achieve the target sample size, improved the specificity of recruitment of individuals with a formal diagnosis of dementia, and facilitated the wider engagement of professional roles such as physiotherapists and occupational therapists. In terms of suitability of outcomes, our reliance on self-reporting regarding physical activity levels permitted individuals to join our study who were already regularly practicing physical activity at higher levels than those desired for our target population and, thus, less likely to benefit from our intervention design. The research purpose of self-reported physical activity levels was as a baseline and follow-up measure, not as a “screening” question. While information sheets can always be improved, there was clear reference to physical activity inclusion criteria. Future research will ensure that eligibility criteria are revised and participants’ activity levels are considered more clearly ahead of them consenting to take part in a study. Finally, our use of RE-AIM was ultimately focused on the planning and conduct of data gathering as opposed to the evaluation phases. While flexibility regarding the use of RE-AIM is both permitted and encouraged, as a research collaboration, our focus from the outset of this work was to better understand uncertainty regarding recruitment, intervention design, and suitability of outcomes following elements of the MRC complex intervention framework. We chose to focus on these elements to provide a clearer narrative throughout.

Comparison With Prior Work and Future Directions

A key finding was that carers of people with dementia were recruited at a slower-than-anticipated rate. In the context of a randomized controlled trial (a future consideration for this work), our recruitment rate of approximately 1 to 2 carers per week through 1 center was higher than medians reported elsewhere [37]. Nonetheless, we did not achieve our intended target of 50 carers of people with dementia, and future work should consider ways to mitigate the risk of slow recruitment, including increasing the length of the recruitment period (ie, beyond 6 months) or considering multisite recruitment over a larger geographical area and broader carer groups. The self-identification and visibility of all carers has been highlighted across a wide array of recent literature [38,39], and other recent work has also highlighted limited uptake of interventions for carers integrated with health care service delivery even in sizable population groups [40]. Future work should consider stronger emphasis on face-to-face conversations, which can be critical for uptake and retention [41]. Despite efforts to add more representation of carers (eg, gender, ethnicity, and geographical setting), inequity in the reach of CareFit was a prominent observation. This may be related to intersecting sources of inequity and disadvantage; for example, a lot of carer research demonstrates a high prevalence of individuals who are female, White, and highly educated [40]. Further qualitative work is required building on foundational knowledge [42-46] as to how we can design preventative approaches to reach those traditionally marginalized in a more equitable manner.

Arguably, one of the most pertinent findings regarding intervention uncertainty was a study dropout rate of 49% (20/41; 21/41, 51% of the participants completed the 8-week study). While this is high in comparison to clinical trials [47], there is systematic review evidence suggesting that high attrition can be common for app-based observational studies [48]. Our additional exploratory analysis of data identified that retention was considerably lower for those carers with high levels of physical activity as opposed to those with lower levels (retention of 5/14, 36% vs 16/27, 59%, respectively). Thus, future work needs to consider how to achieve stricter enforcement of inclusion criteria. CareFit could also be expanded to cater to more stages of change according to the transtheoretical model and signpost such users to other care providers and professionals where required. While our technical issues were few in number, such occurrences are likely highly impactful to carers—both we and others recognize the heterogeneity of motivations present in this population, where personalized delivery is key [7]. Future work would ideally leverage learnings on the digital divide that both we and others [49,50] have noted. For CareFit, this could include reducing the number of download steps for first use, improving the flow of the activity planner (including creating a menu of potential plans), and increasing opportunities to use larger screen sizes such as laptops.

More broadly, facilitators for use across carers and professional stakeholders include integration into the realities of the caring routine and sustaining motivations for physical activity within this context. Many participants highlighted that the app as delivered was fun and motivational and the delivery of short videos (eg, including content for use in 1 minute) in particular could fit into the caring role. While this type of delivery method is relatively novel, technical challenges remained, some relatively addressable (eg, optimization of the planner) but also some more substantial, including those regarding the phone screen size being too small to allow physical activities to be easily conducted. Participants also mentioned that deepening links between the Learn and Activity sections would support carers with content “more about getting me doing stuff.” Analyses also highlighted an appreciation of the fact that human elements are critical to convert autonomy and competency regarding physical activity in carers into long-term maintenance. This includes “championing” of CareFit by other carers—mirroring the importance of social influence across a broader range of behavior change interventions [51]. We hypothesize that future efforts could leverage value both from face-to-face settings (eg, working alongside employers) and from community pages already set up on the app to date, including facilitating more content generation from other carers on a regular (eg, weekly) basis. Stakeholder analysis identified that optimal delivery and support could include a variety of settings, including the use of regular carer assessments, linking more closely with frontline community staff such as pharmacists and those conducting home visits to support adoption and regular use.

Our third key area of interest was the suitability and utility of different outcome measures of the (future) impact of CareFit. Qualitative findings from professional stakeholders suggest no clear consensus on what or how to measure regarding impacts. We conclude that, from a care provider perspective, current economic and staffing pressures mean that cost-effectiveness and reducing demands in care is likely to be a key driver of uptake. Concerns that health is not valued as an outcome it its own right have been raised in other studies [52]. While increasing physical activity levels in marginalized groups could yield significant long-term benefits for individuals and care providers, the short-term impact is critical. Our findings highlight the value in aligning with carer-described goals—something that would align with multiple potential models of carer empowerment [53,54] and that would be of significant interest for future work.

Throughout the work presented in this paper, we relied largely on self-report measures. While the validity of measures such as the IPAQ-SF has been well demonstrated across many different population groups [34], a recent meta-analysis identified low criterion validity for single items, such as single-item sedentary behavior questions [55]. Further development of (single- or multiple-item) measures with greater validity could be of interest to a variety of stakeholders. This could include further exploration of the new outcome measures introduced in this study. With respect to CareFit, measures could take on a dual role as both mediators (advancing the theoretical understanding of CareFit) and creating evidence-based and personalized feedback for carers. Missingness regarding physical activity measures included both activity recording within the app and the use of the IPAQ-SF. A more systematic and focused approach on understanding the recording of physical activity among marginalized groups appears to be of value. We propose that a helpful extension to CareFit (at least for some users) would be to integrate more automated data collection, such as the use of wearables or step counts (eg, collected using mobile phones). Such integration would have to be considered carefully so that those carers with lower health literacy or subject to the digital divide are still able to receive the intervention equitably.

Conclusions

Accessible physical activity for carers of people with dementia remains an unmet societal need. Our work demonstrates that self-recognition and reach of carers of people with dementia are critical. While we cannot currently recommend the progression of CareFit to a randomized controlled trial, both the safety and usability of CareFit were promising. Furthermore, our work builds new knowledge on how to design interventions for carers—particularly the use of extensive video content, which was positively received. We conclude that CareFit sits largely within the context of preventative medicine as many informal carers are facing additional barriers to physical activity compared to noncarers. Further work is warranted to build foundational knowledge on how to further enhance both reach and adherence across key public health behaviors for this critical societal group.