Staff’s involvement in tool development through coproduction activities influenced engagement with the pilot. Managers were aware of the need for staff to feel involved in new interventions:

The CPT was thought to enable more efficient information recording and facilitate coproduced care. Creating notes using the CPT alongside service users facilitated transparency and involvement of service users, and it was quicker than traditional ways of working:

Some participants distinguished between the CPT’s features, which they thought were well designed, useful, and easy to use, and limitations of the tablet computer on which the CPT was hosted:

Issues with internet connectivity and tablet computer log-in problems were a common barrier to using the CPT with service users:

Other barriers included security limitations, and lack of live cross-system communication between the CPT and the host organization’s main electronic patient record (EPR) system. Until secure platforms for information exchange were developed, all data were manually transferred from the CPT system to service users’ EPRs by administrators. The delay in transfer (up to 48 hours) impacted on information available during meetings, sharing of information between providers involved in care, and ultimately how and how often the tool was used:

A recent Care Quality Commission (CQC) report of the mental health organization highlighted the need to improve inclusion of service users’ views in care plans [33]. The CPT could be a key mechanism to improve practices in response to the CQC report’s comments, and Joint Project Board members saw this as an important facilitator to encourage its use within the organization.

In this construct, quotes relating to awareness of service user–specific factors influencing implementation were included. Some participants were guided in their use of the tool by perceived needs and characteristics of individual service users. For example, English language literacy was taken into consideration when deciding to use the tool:

Service users’ levels of self-awareness and stage in their illness influenced CPT use, for example, when service users were not thought to be able or ready to engage in recovery care. Another influencing factor was service users’ attitudes toward technology, with age being a related factor:

When introducing the tool to service users, reasons for declining its use included wanting to “talk to a human being” [Practitioner 03]; seeing it as a wall between themselves and providers; distrust toward technology; and thinking it did not enhance their care experience. Declining to use the tool once, however, did not always exclude use of the tool in subsequent sessions:

The organization already had a focus on recovery-oriented care, including engaging service users in care planning. Participants agreed that the CPT facilitated coproduced recovery-focused care planning by supporting novel and more user-centered conversations with a psychosocial focus:

However, some medically trained staff saw the CPT’s focus on psychosocial aspects of recovery as contradictory to their professional roles, for example, discussing medications:

Organizational IT factors were also raised. The organization was thought to lack adequate IT infrastructure, to be paper heavy, and not incorporating technology within current practice:

The organization and staff were under pressure from increasing numbers of referrals. This provided an incentive for innovations that would compensate for the lack of capacity:

Pressures on the organization impacted on the time and resources staff had available to learn how to use and implement the CPT. Carrying a heavy caseload also shaped capacity for recovery-focused work, as staff had to address the service user’s immediate needs. This influenced perceptions of how the tool could support meetings with service users:

Most participants were positive about the CPT’s ability to aid in coproducing care plans, stating that they “liked the idea of doing a support plan on a device (that was) client facing” [Practitioner 02]. Some participants used most or all of the tool features with different service users, whereas others believed the tool could only be used in certain contexts, for example, guided by service users’ needs. There was reticence among some providers to introduce a tablet computer to the therapeutic relationship:

Issues of confidence and perceived ability to use the CPT were raised by both experienced users of the tool, that is, those trained 6 to 8 months before the interview, as well as inexperienced users, that is, participants trained around 2 months prior. Familiarity with the tool was needed when deciding who to use the CPT with, which components to use, how, and when. It also related to how comfortable staff were in changing practice and introducing new ways of working, potentially influencing the therapeutic relationship:

Staff recruitment to the pilot was led by practitioners from inside the organization, which facilitated getting staff on board. Having continuing user feedback allowed the software developers to make improvements to the CPT software. This meant that some tablet computers needed to be replaced to support updated versions of the CPT, but it also meant that some participants received new tablet computers toward the end of the pilot phase, not allowing enough time to use the new tool in their practice. Restrictions placed by the organization on which service users could be involved in the pilot, that is, individuals who were assessed to be at low/medium risk for experiencing a mental health crisis, also acted as a barrier:

Some participants thought risk should not be a limitation and the CPT could potentially be used with high-risk individuals if enough consideration was given to which aspects of the tool were used, and how it could be used during interactions:

Training provided during implementation helped support participants using the tool. Training included group or one-to-one sessions followed by feedback meetings, but some participants recruited later in the pilot did not always receive similar training:

One participant thought training on the interactive element of the tool was needed to guide integration into practice in a way that does not compromise the therapeutic relationship. Such aspects of using the tool were seen as important because of its objective to support collaborative service provision, coproduced with service users:

Information leaflets explaining what the tool is about and how to introduce the project to service users were thought to be particularly useful.

Aref-Adib et al [14] state the need for better understanding of the contextual and organizational determinants of successful implementation. Our findings highlight the importance of ensuring alignment between external policy, organization and staff priorities, and CPT features to aid intervention engagement. Externally, the CQC inspection findings and increased demand for mental health services provided an incentive for organizational change in ways of working. The CPT facilitated changing practice in ways that met these pressures while aligning with organizational and professional values.

Other highlighted factors include tool adaptability to existing ways of working and attitudes and beliefs toward the digital innovation; stakeholder involvement in the development process is recommended to address these factors and facilitate implementation [14]. In this study, coproduction principles [20] in tool design and development supported engagement in implementation, and the tool was thought to facilitate service user–centered, recovery-focused coproduced care [27], a professional and organizational priority. Uptake was impacted by available organization resources, including IT infrastructure, staff caseloads, and time pressures; staff self-efficacy and knowledge of using the tool; service user attributes; and mobile working–related factors, including internet connectivity and IT system compatibility with the CPT. Findings reiterate the importance of considering such issues early on in digital innovation design [14,36].

Importance attached by providers to the therapeutic relationship when adopting digital health technologies highlights the need to better understand interpersonal aspects of health technologies in clinical contexts [13,18,27,37]. Perceptions of impact on, and concerns about, the therapeutic relationship influenced whether and how the CPT was used, and so did perceptions of relevance to role priorities. Staff assessments of service user needs and characteristics, for example, crisis management, literacy or attitudes toward technology, and uncertainty as to how to most successfully integrate the tool in practice, influenced providers’ choice of who to use the tool with, and how often.

Mental health service users are open to using digital health technologies [5,38], but barriers, such as, intervention complexity [14], can prevent widespread access and use among individuals with increased needs, for example, those experiencing psychotic symptoms or learning difficulties. At the same time, low IT and health literacy and digital inequalities among individuals with mental illness also impact on innovation uptake [14,39-41]. Our findings suggest in some cases, staff’s perceptions of service user characteristics, for example, literacy skills, may result in some service users being excluded from interactive health technologies in a care setting, but more research is needed to better understand staff decision making on this aspect, including differences in perspectives between providers from different professional backgrounds. Views of service users from underrepresented groups at risk of digital inequalities should also be explored [41].

Training and access to continuing support on technical and interactive aspects of the intervention during and after implementation may enhance efforts to integrate technology into routine practice [14,42,43]. Training that approaches use of digital innovations in a more reflexive and critical way [26] might address skepticism toward the innovation and concerns over its impact on the therapeutic relationship. When planning such activities, workload pressures and time available for staff to attend training need to be considered [14]. Training informed by action research can result in changing practice [44], and its usefulness in promoting acceptance of health technologies should be explored [45], especially because of its philosophical similarities with principles of coproduction [20].

Theoretical grounding of implementation research allows for conclusions to be drawn as to the relevance of findings to other settings and contexts, allows comparisons, and guides further research [34]. In our study, the CFIR informed data analysis and identified factors shaping intervention implementation, following examples supporting its use in qualitative research evaluations [30,34,35]. The CFIR was useful in guiding categorization of factors and capturing overarching implementation factors involved in the CPT’s uptake. Using the framework in the analysis stage presented challenges in assigning data items to individual dimensions or constructs, and identifying which ones were the most salient in our data, an issue already raised in the literature [10]. In our case, challenges reflected the unique nature of the CPT as a tool used by both providers and service users simultaneously in the context of a care meeting.

Analytical ambiguity existed between the categories Patient needs and resources and Individual characteristics when coding service user–related factors, as both were end users of the tool in the same context and setting; the tool’s dynamic and interactive nature also made difficult distinctions between, for example, the intervention’s adaptability, complexity, and design quality when categorizing data items. With more treatment interventions provided in an interactive way through mobile technology, it is essential for frameworks such as the CFIR to capture the dynamic and multidimensional nature of technological interventions [10]. The complex nature of such interventions, its impact on implementation, and the CFIR’s ability to adapt to and capture this complexity should be further explored in future evaluations of such mobile mental health care interventions.

Varsi et al [46] discussed the broadness of the CFIR that can restrict one study’s ability to capture the big picture represented in the framework, without explicitly addressing all dimensions during data collection [34]. This can be because of time limitations restricting researchers’ ability to explore all constructs within a single interview, but also recruitment limitations, when stakeholders that might represent different views are not included in the sample [46]. One way our study tackled this was to interview both practitioners using the CPT and senior managers who had a broader perspective on implementing technological innovations.