The number of people being diagnosed with cancer is continuing to increase in the United Kingdom, with an estimated 4 million adults living with and beyond cancer (LWBC) by 2030 [1]. Many people LWBC experience challenges related to cancer and its treatment, including increased fatigue, pain, psychological distress, and reduced physical capacity [2-5]. These challenges can significantly impact their quality of life and well-being [6]. Therefore, the importance of improving outcomes for those LWBC is vital [7]. A large body of trial data demonstrates that physical activity (PA) improves many outcomes after a cancer diagnosis, including reduced fatigue, pain, anxiety, depression, and sleep problems and an overall improvement in quality of life [8-11]. Observational data suggest that PA is also associated with improvements in survival [12-14]. In light of this ever-growing evidence base, the World Cancer Research Fund recommends that people LWBC follow the guidelines for healthy populations in achieving at least 150 minutes of moderate-to-vigorous PA (MVPA) per week and recommend limiting the amount of time spent sedentary [15]. Despite this, many people LWBC are physically inactive, with Macmillan Cancer Support (United Kingdom) estimating this to be as high as 80% of those LWBC not meeting recommended PA levels [16]. This is supported by the systematic review of 41 studies conducted by Wong et al [17] that indicated that only around a third of people LWBC were meeting PA guidelines, although this ranged from 16% to 88% across studies. Brisk walking is a form of MVPA that may be more appropriate for people LWBC due to its accessibility and achievability. In a systematic scoping review of 98 studies in people LWBC, walking was cited as the preferred type of PA across all cancer types and treatment stages [18].

APPROACH is an app-based, multicomponent intervention informed by extensive development work with individuals with breast, prostate, or colorectal cancer and cancer nurse specialists [19,20]. It focuses on promoting and monitoring brisk walking using a publicly available mobile phone app, alongside brief behavioral support in the form of a specially designed leaflet, walking planner cards, and 2 phone or video calls with a trained researcher (CB) [21]. In line with the Medical Research Council framework for intervention development and evaluation, the pilot trial explored the feasibility and acceptability of conducting a complex PA intervention trial with people LWBC. A total of 90 people diagnosed with breast, prostate, or colorectal cancer were recruited for the pilot randomized controlled trial (RCT), with 49% (n=44) randomly assigned to the intervention group and 51% (n=46) to the control group. APPROACH pilot results demonstrated a high retention rate (97%) and high assessment completion rates (>86%), indicating that the trial procedures were feasible and acceptable to be carried out as intended in a confirmatory, phase-3, larger trial [22]. In addition, results showed that the intervention was delivered successfully with 98% receiving at least 1 behavioral support call and 95% of participants downloading the app [22].

The importance of conducting process evaluations within RCTs has been emphasized to explore the way in which any complex intervention is implemented [23,24]. This can help uncover why interventions are successful or unsuccessful, determine why they may have unexpected consequences, as well as explore how an intervention that is effective can be optimized [23]. Moore et al [25] provided specific guidance for carrying out process evaluations, which highlighted the importance of exploring the implementation (eg, the fidelity of intervention calls), the mechanisms of impact (eg, views on the different components of the intervention), and the contextual factors influencing use and outcomes (eg, barriers or facilitators to engagement) [25].

Therefore, this paper extends the published APPROACH feasibility results [22] with the following aims: (1) to evaluate the implementation of the APPROACH intervention and the fidelity of the delivery of intended behavior change techniques (BCTs), (2) to identify the potential mechanisms of impact that underlie behavior changes attributed to the APPROACH intervention, and (3) to better understand how contextual factors influence engagement with the APPROACH intervention.

This was an embedded design, mixed methods study where qualitative and quantitative data were collected simultaneously with equal priority [26]. Data were collected as part of the APPROACH pilot RCT [21,22]. The pilot RCT compared an app-based, brisk walking intervention delivered alongside usual care with a control arm (usual care alone) in people diagnosed with breast, prostate, or colorectal cancer at a single hospital site. The trial was registered on the International Standard Randomized Controlled Trial Number registry on April 16, 2021 (ISRCTN 18063498). The primary outcome for the pilot and future RCT is weekly minutes spent brisk walking (a cadence of 100 steps per min or more [27) measured by an activPAL accelerometer (PAL Technologies Ltd). Following baseline assessments, participants were individually randomly assigned (1:1 allocation) using minimization to either the control or intervention arm, stratified by cancer type (breast, prostate, or colorectal) and disease status (metastatic vs not).

The pilot RCT included 90 participants: 49% (n=44) in the intervention arm and 51% (n=46) in the control arm. All participants had a confirmed diagnosis of breast, prostate, or colorectal cancer (localized or metastatic). At the point of screening, localized participants had to be within 6 months of completion of radical treatment. This criterion was not applied to participants with metastatic disease. All participants required a clinician’s sign-off that their life expectancy was >6 months. All participants self-reported achieving <150 minutes of MVPA weekly. Full inclusion and exclusion criteria have been published previously [21].

The recruitment procedure and trial data collection have previously been described [21,22]. Participants completed assessments at baseline (time point 0 [T0]) and 3 months (time point 1 [T1]; operationalized as 12-16 wk after randomization). The intervention included an endorsement letter from a member of the participant’s clinical team, alongside the provision of a leaflet with information about the importance of PA after cancer and a recommendation and instructions on how to download the freely available NHS Active 10 app. The NHS Active 10 app promotes brisk walking in bouts of 10 minutes, called “Active 10s.” This was augmented with 2 behavioral support phone or video calls with the intervention facilitator (CB). The app was chosen after previous qualitative work with people LWBC, and clinicians identified key features that were important to be offered within the app and highlighted the importance of the app being supported by a professional organization, such as the NHS [19,20]. These behavioral support calls were underpinned by habit theory [28] and BCTs shown to be effective in promoting PA [29-31] and involved supporting participants in downloading and using the app and discussions around setting PA goals. The intervention facilitator was trained in the principles of behavior change theories and in the application of BCTs [32], with a thorough understanding of habit theory [33]. This training, alongside previous experience in delivering health behavior change interventions, allowed them to conduct conversations with patients closely mirroring those a health care professional might have with a patient in a routine care situation. The intervention calls took place via Zoom (Zoom Video Communications) [34] or telephone and were recorded by the intervention facilitator with the participants’ permission. The first call took place approximately 1 week after randomization, and the second call took place approximately 4 weeks after the first call to check in with the participant about their goals and recap any information required. Participants were asked during the first intervention call if they had downloaded the NHS Active 10 app, and the intervention facilitator noted this in their records. Participants were also given 12 copies of a walking planner card that was designed to enable them to plan how many “Active 10s” they were aiming for and how they were going to achieve these, including where and when they would complete them.

This pilot study was approved by the Yorkshire & The Humber-South Yorkshire Research Ethics Committee (21/YH/0029, Health Research Authority and the local hospital. All participants gave informed consent and the data reported are anonymized.

Table 1 presents the sample characteristics of the 44 participants in the intervention arm. Most (42/44, 95%) of the participants were of White ethnicity, comprising an equal number of male participants (22/44, 50%) and female participants (22/44, 50%), with a mean age of 63 (SD 11; range 40-85) years. Participants had received a diagnosis of breast cancer (18/44, 41%), prostate cancer (18/44, 41%), or colorectal (8/44, 18%) cancer. The CONSORT (Consolidated Standards of Reporting Trials) diagram and the flow of participants through the study have previously been reported [22]. After eligibility screening and assessment of interest in taking part, the study information sheet was sent to 148 patients, and 63% (n=93) consented to participate, with 61% (n=90) being randomized.

Figure 1 presents the theme diagram showing the identified mechanisms of impact and the contextual barriers and facilitators that affect these mechanisms and intervention engagement.

This study combined data collected as part of the APPROACH pilot RCT to assess the implementation of a multicomponent, app-based behavioral intervention to promote brisk walking in people LWBC. The findings of this process evaluation demonstrated proficient implementation of the intervention and suggest that there are several mechanisms of impact underlying the efficacy of the intervention as well as contextual factors that can be barriers or facilitators to engagement.

The successful delivery of intervention components and intended BCTs is essential for attributing any changes in behavior to the intervention in question [41]. This study demonstrated proficient implementation, with most participants reporting engagement with the intervention components, including downloading the app, reading the leaflet, receiving the behavioral support call, and using the planners. In addition, fidelity is rarely reported in evaluations of PA interventions [29]. In a systematic review of 21 studies assessing the quality of measuring delivery fidelity in PA interventions, Lambert et al [42] reported considerable heterogeneity in evaluating delivery fidelity. This study reports high delivery fidelity of the intended BCTs in the behavioral support calls [21]. While interventions with only a single intervention facilitator often exhibit higher fidelity, investigating how various intervention facilitators across different contextual settings deliver BCTs is essential for understanding the real-world application and scalability of interventions [43]. Particularly where multiple intervention facilitators are involved in the delivery, future research should consider factors such as personal characteristics and individual context when evaluating intervention delivery [44,45].

The BCTs called information on health consequences, action planning, and habit formation showed the highest fidelity, and these BCTs appear to influence key mechanisms of impact highlighted by participants in the qualitative interviews. Empowering participants with information on why they should go walking and providing them with information about its benefits appeared to enhance their engagement and adherence to their planned walks. The delivery of and engagement with the action planning BCT is promising, as this BCT was shown to be associated with larger effect sizes in a systematic review of 26 studies evaluating BCT effectiveness in PA interventions in adults who are healthy and inactive [29]. This study extends this finding and suggests that this BCT is suitable and appropriate for people LWBC and should be included in future intervention designs with this population. The BCTs that were not successfully delivered included those called nonspecific reward, nonspecific incentive, and framing or reframing. Despite the BCT called nonspecific reward receiving a low fidelity rating in the recorded behavioral support calls, the qualitative interviews suggested that this BCT was effectively covered in other aspects of the intervention, with participants reporting enjoying working toward the rewards and trophies in the NHS Active 10 app. The low delivery fidelity of the BCT called framing/reframing could be attributed to participants’ voluntary enrollment in a PA trial. It is likely that they already recognized the importance of PA even without fully recognizing its role in life beyond a cancer diagnosis [46,47].

The mechanisms of impact identified by participants in the qualitative interviews and follow-up questionnaire were in line with previous research. For instance, self-monitoring of behavior is one of the most frequently used components in complex PA interventions [48], and multiple systematic reviews have demonstrated its effectiveness in increasing PA that is maintained long-term [49-51]. In this intervention, the NHS Active 10 app allowed participants to track their brisk walking and total walking, and participants reported that being able to see their improvement over time motivated them to continue their behaviors. In a similar pilot RCT using a self-monitoring app alone, Ormel et al [52] reported that participants in the intervention group increased their activity more from baseline to 6 weeks, but this difference was not maintained at 12 weeks. The authors attributed this to a potential loss of novelty and interest in the app. Although not powered to detect differences in PA, the findings of this study suggest that the second behavioral support call was important in continuing encouragement of monitoring of behavior and reviewing of goals, with participants describing a sense of accountability with the later call. This feedback indicates that a light-touch intervention call to check in with participants can be beneficial in consolidating commitment to PA goals in people LWBC. While technology can help reduce resource demands, future research should consider the benefits of a low-burden behavioral support call to augment an app as an already powerful and scalable intervention component.

Participants extensively discussed the app as the driving component of the intervention, supporting our previous reports of high engagement with the NHS Active 10 app [22]. The discussions revealed that another mechanism underlying behavior change in this context was the ability of the app to increase motivation toward reward using gamification techniques [53]. The performance of the desired behavior (brisk walking) was reinforced by the positive feelings of encouragement and dedication resulting from these cups and trophies in the app.

The APPROACH intervention was informed by habit theory [21], and the identification of habit formation as a mechanism of impact in both the qualitative interviews and questionnaire data demonstrates that this BCT was delivered effectively in the intervention. Participants reported that the leaflet, app, behavioral support call, and planners helped to establish sustainable walking habits. Gardner et al [54] define habit formation as “learning cue–behaviour associations, that when cued, automatically generate action impulses.” This sense of automaticity was described by participants in the qualitative interviews, whereby consistent repetition of their walking daily led to the enactment of brisk walking. This meant that they engaged in brisk walking even during activities that previously would not have involved this exercise intensity. These findings are also supported by the increased SRBAI results from baseline to follow-up, which showed an increase in the initiation of walks as well as the way walking was executed (ie, briskly) [55]. The importance of encouraging long-term engagement in PA after the intervention period is highlighted in systematic reviews of PA maintenance in cancer populations that report only modest improvements at longer follow-up time points [30,56]. The results of this study endorse the integration of habit theory into future interventions aimed at increasing PA to overcome the challenge of sustaining behavioral changes over time [56].

To reinforce the idea of habit, the walking planner cards were designed to promote habit formation, facilitate planning, and prompt participants to engage in PA [21,54]. However, both questionnaire data and interview data showed that the structured design of the planner cards was not compatible with the day-to-day changing schedules and lifestyles of some participants, highlighting the importance of conducting this process evaluation to account for and reconsider this contextual aspect of the intervention and future similar interventions. The need for flexibility in lifestyle and planning was reported as a barrier by participants and further confirmed by the reflections of the intervention facilitator (CB) after discussing their use with participants.

Other contextual barriers included the impact of cancer and its treatment, having other competing commitments, technical difficulties with the app, and environmental influences. The side effects of cancer and its treatment have previously been identified as a key barrier to PA participation in systematic reviews [18] as well as in our own preparatory work for this pilot RCT [19]. Participants in this study reported that the impact of cancer and its treatment inhibited their ability to engage in some elements of the intervention due to different physiological, structural, and psychological factors. Cancer-related fatigue is the most reported symptom in people LWBC who have undergone treatment with prevalence estimates of up to 90% of those treated with radiotherapy and 80% of those treated with chemotherapy [57,58]. In this study, participants discussed fatigue symptoms and felt that engaging with the intervention during treatment would have been difficult. Beyond this physiological barrier, the structural barrier of having any appointments for their cancer care also reduced their ability to engage with some components, including the planner card, as there were many hospital appointments that they had to attend and plan around. We have previously reported on the perceived suitability of the timing of the APPROACH intervention, with most participants feeling that it was reasonable [22]. While some participants felt that engaging during treatment would be difficult, others felt that it was useful to have something else to focus on and have control over [22]. Previous reviews in this area have also reported discrepancies in the preferred timing of PA intervention delivery within the cancer care pathway [18,59]. Involvement in PA at an earlier stage has been associated with improved treatment response, tolerance, and quality of life [13,60,61]. Considering, the delivery of the APPROACH intervention during and after the treatment for cancer is still endorsed while recognizing the contextual barriers, such as this, when interpreting APPROACH intervention outcomes. These findings highlight the importance of involving patient perspectives in future intervention design with this population, with the acknowledgment that different stages of the treatment pathway can facilitate or inhibit PA participation and should be accounted for when assessing intervention delivery and engagement.

Facilitators to engagement included having support from others, trust in health care professionals, the perceived usefulness of the intervention, and the perceived benefits of engagement with the intervention. The BCTs called social support (practical) and social support (emotional) were competently delivered in line with the protocol [21,31]. Accordingly, participants recognized how having support from others, such as family members and partners, enhanced their engagement with the intervention and how they felt supported by the intervention facilitator. Social support has previously been identified as highly important in PA engagement [62], with reasons, such as accountability, being cited as helping to facilitate and promote engagement [63]. In our preparatory work for this RCT, trust in health care professionals emerged as a crucial factor influencing engagement [19], and this was highlighted again in this study, where incorporating an endorsement letter from the clinical care team enhanced app credibility.

Limitations of this study include that only 1 intervention facilitator delivered all behavioral support calls, which may explain the high-fidelity ratings. It is crucial to consider the transferability of the intervention across individuals, particularly when envisioning integration into routine NHS care [20,22]. In addition, there may be some recall bias influencing results, as most of the questionnaire data were collected at the 3-month follow-up [64]. Participants may have had difficulty in answering questions on earlier components (eg, leaflet information) compared to components they were still using (eg, app). Due to the pilot nature of this study, where feasibility and acceptability were the main outcomes of interest, we were unable to examine how engagement with each intervention component influenced the primary outcome of brisk walking. For future research, the application of the Multiphase Optimization Strategy with a factorial design could offer more insights into how differing engagement with each component can impact the main outcomes and help inform intervention optimization for larger efficacy RCTs [65,66]. Finally, despite its recognition as a potential mechanism, it is difficult to assess how the intervention helps to establish longer-term habits, which are key for PA maintenance [67], as this pilot RCT only examined outcomes at 3 months.

This study extends our previously published findings on the APPROACH pilot RCT [22] by demonstrating that the intervention was delivered as intended with high levels of engagement from participants. In addition, this paper highlighted the potential mechanisms through which change occurs, such as habit formation and behavioral monitoring, which are in line with the intended BCTs used in this intervention. The process evaluation also highlighted important contextual factors to consider when progressing to the APPROACH main trial, including facilitators, such as social support, which played a significant role in promoting adherence to the intervention. The protocol for the definitive RCT will report on adaptations made to APPROACH based on the feedback gathered in this study. This process evaluation provides strong support for the progression to the stage-3, definitive RCT to evaluate the effectiveness of the APPROACH intervention (began in November 2023) and enables a more nuanced understanding of how the APPROACH intervention works and the contextual factors to consider with implementation.