Recent technological advancements have led to the proliferation of assistive technologies (ATs), including assistive robotic arms, improving the quality of life for individuals with spinal cord injury (SCI) [1]. SCI is a profound, life-altering condition that affects approximately 15.4 million people worldwide [2], with significant impacts on individuals, caregivers, and health care systems [3-6]. Tetraplegia, the partial or complete loss of motor and sensory function in all 4 extremities and the trunk, is a common manifestation of cervical SCI [7]. Upper extremity function ranks highest in importance among individuals with tetraplegia, often prioritized over functions such as bowel, bladder, sexual, or walking abilities [7]. Disabled upper extremities not only diminish independence and social participation but hinder employment prospects, leading to increased care costs [8,9]. Comparatively, individuals with SCI often experience lower quality of life levels than their able-bodied peers [10,11], and regardless of the specific nature of their impairments, they frequently depend on caregivers to meet their daily needs [12,13]. Paid caregiving services can impose a significant financial burden on either the individual or the funding agency [14], whereas unpaid caregiving, provided mainly by family members, results in significant psychological implications [11,15,16].

Persons with impaired upper extremities, whether due to tetraplegia or other causes, such as multiple sclerosis or cerebral palsy [17], heavily depend on ATs to enhance independence and participation in activities of daily living (ADLs) [18,19]. ATs augment individuals’ capabilities and broaden the range of activities they can engage in. Commonly used adaptations include cutlery, environmental control systems, writing orthoses [20], wheelchairs, prosthetics, and orthotics [21]. The application of AT not only aids the user but also indirectly benefits caregivers, notably family caregivers. Atoyebi et al [22] emphasize the potential of AT in reducing caregiver burden and alleviating stress. Specifically assistive robots, as defined by Kyrarini et al [17], serve as compensatory ATs and offer external assistance controlled by the user through an interface [23,24]. They facilitate a wide range of activities for persons with upper extremity disabilities, enhance care and promote independence across different areas [25,26]. One example is the Functional Robot with Dexterous Arm and User Friendly Interface for Disabled People (FRIEND) system, a wheelchair-mounted robotic manipulator designed to assist users with tetraplegia in tasks such as drinking and eating. The FRIEND IV, the latest iteration, integrates a 7-degree of freedom robotic arm and a 2-finger gripper and has been tested in a use-scenario, achieving a 95% success rate for cataloging books (of a performed 100 runs) with the end user’s intervention [27]. The Jaco 2 robotic arm (Kinova Inc), a widely used commercially available system, offers 6- or 7-degree of freedom configurations and can be equipped with 2- or 3-finger grippers. It has been applied in various contexts, such as adaptive feeding systems that integrate force sensors and cameras to detect and manipulate food for users. In an efficacy study with 31 participants, the Jaco robotic arm has demonstrated the potential to reduce caregiving time by approximately 41% [28]. Beyond feeding and drinking tasks, assistive robotic systems have also been implemented in other ADLs. The Baxter humanoid robot, for example, has been used to provide personalized dressing assistance, using force-minimizing methods to help users with upper-body impairments wear sleeveless jackets [29]. Other experimental systems have explored applications such as beard shaving, hair brushing, and bathing, although these tasks remain in early stages of development and require further research to ensure safety and usability [17]. However, most robotic systems are still in the research phase and require significant customization and personalization to effectively assist persons with impaired upper extremity in their daily life.

There have been efforts to promote high-performance AT as prototypes, such as international contests like the CYBATHLON hosted by the Swiss Federal Institute of Technology Zurich in Switzerland [30,31]. Nonetheless, the need [32,33], the transfer, and the adaptation of AT in everyday life remain a transnational challenge [1,34,35], and many robotic innovations fail to reach the market [36]. This discrepancy can be attributed to various factors, including a focus on technical feasibility over user needs, complexity, cost, and the need for technical assistance for use [36,37]. For effective market integration, factors such as reliability, cost-effectiveness, esthetics, functionality, and usability must be carefully considered, with user-centered design principles playing a relevant role [23,38,39].

The need of structured knowledge transfer, reimbursement strategies, and the assessment of the readiness of the AT market to successfully transition a prototype into a market-ready product [30] can be met with implementation sciences, offering valuable methods for the systematic uptake of evidence-based practices [40]. The implementation sciences acknowledge the relevant influence of economic, political, professional, and sociohistorical context factors on the adoption of evidence in practice [40]. The aim of this study is to investigate and reduce the research translation gap of implementing ATs into social and health insurance systems, with a specific focus on integrating a user-centered designed assistive robotic arm into the everyday private lives of individuals with tetraplegia in Switzerland [41]. The study’s objectives are (1) identifying and involving relevant stakeholders, (2) assessing contextual factors (barriers and facilitators), and (3) proposing an implementation strategy for the robotic arm.

A qualitative design was applied, conducting 3 online focus group interviews with representatives from 8 stakeholder groups relevant to the robotic arm’s implementation, aiming to identify barriers and facilitators. This study was performed from a constructivist or interpretivist point of view using a general deductive approach. This research was reported in accordance with the Standards for Reporting Qualitative Research by O’Brien et al [42].

The Consolidated Framework for Implementation Research (CFIR) serves as a widely used guide in implementation research, comprising 39 constructs synthesizing existing theories applicable across all implementation phases, allowing researchers to select the most relevant constructs for their specific study setting [43-45]. It aims to identify and explain factors influencing implementation outcomes, consisting of 5 major domains: innovation, outer setting, inner setting, individuals involved, and implementation process [45]. The innovation is related to the characteristics of the innovation being implemented into a particular setting. Outer and inner settings represent a dynamic interface, where the outer setting generally includes the political, social, and economic context, such as community and state factors, while the inner setting comprises the structural and cultural context. Individuals involved represents the individuals involved with the innovation and the implementation process. In this domain, cultural, organizational, and professional aspects as well as individual mindsets and norms are considered. The implementation process includes the development needed during implementation. According to Damschroder et al [46], researchers should choose the most relevant constructs from the 39 offered in the CFIR to suit their specific study context. These constructs can guide diagnostic evaluations of the implementation environment, monitor implementation progress, and interpret findings in the research studies or qualitative improvement initiatives.

The robotic arm at the center of the focus groups was a prototype developed by the Bern University of Applied Sciences in Switzerland. It competed in the Assistance Robot Race at the CYBATHLON 2024, mastering 10 different tasks of daily life, such as removing a parcel from the mailbox or navigating a touch screen. However, it has not yet been subjected to trials beyond this competition [47]. The robotic arm system consists of a Kinova Gen3 robotic arm with 7 joints, which can be mounted on an electric wheelchair. A 3D camera positioned near the gripper provides visual input for the current task. The vision system uses machine learning algorithms to identify objects, enabling the robot to approach and grasp them in a semiautomated manner, thereby saving time in completing the task. A graphical user interface displayed on a tablet near the user presents live and processed images from the camera. Users with some arm mobility can interact with the system by using the tablet’s touch screen to issue commands, while those without sufficient mobility rely on speech recognition powered by Whisper (OpenAI) [48]. The robotic arm system is designed to ensure simple operation and intelligent robotic behavior. The architecture of the system itself is independent of the robotic arm, allowing for flexibility in implementation. For instance, while the current configuration uses the Kinova Gen3 robotic arm, it could be replaced with an alternative model, such as the BATEO (ACCREA Engineering) arm, without compromising the system’s functionality. In addition to the existing literature on different robotic arm systems, a cross-sectional study focusing on the identification of needs for ATs for upper limbs in persons with tetraplegia was considered for the design of the robotic arm (N Hutmacher, unpublished data, March 2025). An overview of the robotic arm system can be found in Multimedia Appendix 1.

Representatives from 8 stakeholder groups (affected persons, as well as those who could provide perspectives on engineering or technology, legal matters, nursing or care, therapy, social counseling, social insurance, and political considerations) participated in the focus group interviews. The nursing or care stakeholder group included professional caregivers and family members who provided caregiving, the therapy group comprised occupational and physical therapists, and the social insurance group included representatives from various Swiss social insurance organizations (Swiss social insurers according to Federal Accident Insurance [UVG], Federal Military Insurance [MVG], and Federal Law on Invalidity Insurance [IVG]). Inclusion criteria required stakeholders to have specific professional roles or specialized knowledge, be aged >18 years, and have a good understanding of German.

Stakeholders were organized into focus groups, each consisting of 6 to 9 participants, with a minimum of 4 stakeholder groups represented in each focus group. This approach aimed to facilitate discussions and ensure diverse perspectives were represented among stakeholders in a heterogenous group composition [46,47].

Purposive sampling aimed to achieve maximum variation in specialty by using a snowball approach to contact official representatives from insurance institutions, government offices, medical clinics specializing in SCI, assistive product retailers, and personal contacts. Recruiting involved contact via email or phone by the main researcher [42].

The researchers found consistent discussion of similar barriers and facilitators after 3 focus group interviews, confirming data saturation.

To avoid increased time expenditure with long journeys, the 120-minute focus group interviews were conducted online via Microsoft Teams during January 2024 to February 2024.

Data collected for the general implementation outlook were informed by extensive literature research.

Data collection followed a semistructured interview guide (Multimedia Appendix 2) developed by the research team based on literature review. The interview guide was refined in 3 one-on-one online interviews with participants (including a representative of engineering or technology, legal perspective, and social insurers) who did not participate in the focus groups. The guide was supplemented by the discussion and answers from the participants in the one-on-one meetings. The final guide contained the areas first impressions, opportunities, barriers, dealbreakers, and outlook or solutions.

Ground rules, a schedule, and a photograph of the robotic arm as stimulus material (Multimedia Appendix 3) were introduced at the start of each focus group interview. Participants identified themselves by role and name, with sessions being video and audio recorded using the Microsoft Teams application. Moderation was handled by the main researcher, and note-taking was handled by a note-taker.

A total of 3 focus group interviews were conducted with 6 to 9 representatives of the different stakeholder groups shown in Table 1 (focus group 1: 9/22, 41%; focus group 2: 6/22, 27%; and focus group 3: 7/22, 32%). A total of 7 stakeholder groups were represented in at least 1 focus group; no participant engaged in more than 1 group. Family caregivers from the nursing or care and political perspective groups did not participate in the study. Most study participants (13/22, 59%) were female. Participants appeared to be highly experienced in their professional roles; they had worked on average 15 (SD 9.23) years in the current professional function.

The Focus Group Illustration Mapping tool, as outlined by Pelz et al [49], facilitated a comprehensive summary of focus group interview content by combining survey methods with content evaluation. It used the semistructured interview guide as its framework, with key aspects first impressions, opportunities, barriers, dealbreakers, and outlook or solutions serving as central nodes. Relationships between concepts were represented using connecting arrows, with color codes distinguishing stakeholder groups. Visualizations were created using the web-based tool Miro [50], and PDF versions were sent to participants for consensual validation (Multimedia Appendices 4-6). The illustration maps were translated from German to English (Multimedia Appendices 7-9).

The CFIR constructs were selected after intensive examination of all 39 constructs according to the time of the study. A deductive approach was used to code data from the illustration maps, which were aligned with domains 1 to 4 of the CFIR framework.

For the implementation outlook, a similar deductive approach was used to code data extracted from literature research, focusing exclusively on the implementation process domain.

To ensure credibility of the findings, participant validation was used. A link to a Microsoft Forms survey allowed the validation of the illustration map by each participant anonymously, where either the option “I agree with the visual summary” or the option “I do not agree with the visual summary,” followed by an open text field for feedback, could be chosen (Multimedia Appendices 10 and 11). The response rate for the validation was 77%, with 17 participants completing the survey. The illustration maps were translated from German to English using 3 translation programs: DeepL (DeepL SE), LEO (LEO GmbH), and Dict.cc (dict.cc GmbH). Data triangulation was achieved through semistructured interviews and document analysis of online research, converging multiple data sources to validate identified themes and patterns [42].

Written informed consent was obtained from all participants outlining the investigation’s purpose, the use of data, and their publication (Multimedia Appendix 12). Given the design of the study and the nature of the data collected, formal ethics approval was not required. According to Risk Category A as defined in Article 7 of the Swiss Human Research Ordinance (HRO), studies involving noninvasive procedures and fully anonymized qualitative data, such as focus group interviews, are exempt from requiring ethics approval [51]. The research was conducted in accordance with ethical principles respecting the dignity, privacy, and autonomy of the participants.

The implementation outlook was structured on the fifth domain of the CFIR (implementation process) displayed in Table 2; the included data came from literature research. The formation of an experienced implementation team is proposed to lead the implementation [40], while the needs for ATs for individuals with tetraplegia have been identified in a cross-sectional study (N Hutmacher, unpublished data, March 2025). As the project progresses, insights will be mapped to CFIR constructs, and an implementation plan will be developed, including pilot testing. The use of hybrid type 1 and type 2 designs will allow to assess both effectiveness and implementation within real-world contexts [53].

Throughout all focus group interviews, there was agreement that the present stakeholder groups were relevant to the implementation of the robotic arm. A stakeholder group that was frequently discussed among the interviews was individuals providing a political perspective, which is also emphasized by Wolfenden et al [54]. According to a recent study by MacLachlan et al [56], AT policy should be grounded in evidence, but its foundational framework must go beyond rigid scientific criteria and be more inclusive. This is an important consideration for governments because the capacity to implement can be hindered by the volume of concurrent initiatives, potentially leading to low prioritization of the implementation, which could cause the implementation to fail [57]. In this context, stakeholder perspectives, contextual factors, cultural nuances, available resources, and systemic viewpoints should be systematically evaluated and synthesized in a transparent manner to enhance their credibility and inform policy decisions. MacLachlan et al [56] even propose the role of policy entrepreneurs, translating technical content into compelling information to engage politicians, to network and interact with the key stakeholders. Involving the current wheelchair-bound Swiss national councilor, as discussed in the focus group interviews, could facilitate cross-sector collaboration and have positive effects on the implementation of the robotic arm [58].

All stakeholder groups agreed that the innovation addresses the needs of the recipients, specifically the desire for greater autonomy among individuals with tetraplegia. The robotic arm would allow for improvement in daily life participation, and although some activities may still be challenging or time consuming [59], addressing the impairment of the upper extremities and the autonomy provided was valued by the users [7]. Because the robotic arm may not allow complete independence for affected persons, partial or continuous assistance from a caregiver might still be needed, which was perceived as negative in the focus group interviews. Similar results were found by Gelderblom et al [24] assessing the MANUS robotic arm, where the continued dependence on caregivers was perceived as stressful to the user. Nevertheless, according to the focus group interviews, users’ quality of life was expected to improve with the use of the robotic arm. This is reflected in the observations of Maheu et al [28], where the empowerment of the user enables an improvement in the quality of life, not only of the user but also of the caregiver. This facilitator must be emphasized as an important contribution of the robotic arm [60]. Specific assessment of how ATs, more precisely robotic arms, influences the quality of life of affected persons and of caregiver needs to be addressed in future research.

The complexity of the requirements of the innovation and the need for individuality with each user were perceived as barriers to the implementation of the robotic arm. The involvement of users in the development and implementation process was emphasized across the focus group interviews. This not only provides valuable insights that can enhance the effectiveness of the innovation [61,62] but also leads to higher user satisfaction. According to Martin et al [63], users that possess a greater understanding of ATs and feel more informed tend to exhibit higher satisfaction with the devices. Likewise, when their requirements are unmet, the users tend to express less satisfaction levels with the technology. It is anticipated that if usability limitations, such as use efficiency and overall satisfaction, remain unaddressed, users are prone to abandon their AT shortly after acquiring it [30,64]. This is an important factor to consider when assessing the readiness of the AT market and the implementation of the robotic arm moving forward. Furthermore, piloting, as mentioned in the focus group interviews, would allow users to gain experience and further adapt the innovation based on their feedback [65].

According to Nicolet et al [66], it is crucial to involve not only users but also other diverse stakeholders in the early stages. This facilitates the dissemination and transfer of findings, while also enhancing credibility and acceptance within the population. This is relevant as the implementation of the robotic arm could be further facilitated by support from the population. The general positive attitude among the focus groups toward technology as the future presents an optimal prerequisite for this aspect. The favorable view of robots was also shown among the older adult population in Switzerland, although concerns about personal connection and data privacy were raised, supporting similar statements from the focus groups [67]. Understanding patient needs and resources and effectively prioritizing them in combination with the involvement of the public, strongly correlate with successful implementation and are seen as a way to secure the acceptance and implementation of policies [57,68]. These are relevant aspects to consider as we move forward with the implementation of AT, specifically the robotic arm.

A relevant barrier discussed across all the focus group interviews was the financing of the robotic arm. Its current costs as a prototype exceed individual financial resources of affected persons, making it unaffordable. In Switzerland, the governmental social or health insurance system is expected to reimburse costs for ATs. The primary focus of social insurers lies in the cost-effectiveness of the assistance they are considering reimbursing, which according to Romer et al [69] specifically refers to the extent of cost savings achievable through the acquisition of the AT in relation to the overall cost of care for the user (cost of labor of the caregiver and cost of technical aids), which the robotic arm could replace. Additional indirect economic advances of the robotic arm encompass various benefits, such as reductions in unemployment compensation expenses, as users regain employment opportunities or increased workload can be allowed or area of application of work be expanded [69]. According to Bensi et al [56], cost savings of €290 (US $319) per individual over a span of 5 years resulted from enhanced autonomy through ATs and decreased reliance on personal assistants. The delicate balance between affordable supply of needed assistive products and high quality to foster social benefits alongside financial profitability remains a challenge. However, evidence on the effectiveness and cost-effectiveness of assistive robotic arms is limited [70], necessitating further research contributing to the implementation process [71].

In addition, determining the cost bearer within Switzerland’s social insurance system is crucial for the implementation of ATs and was mentioned as a barrier across the focus group interviews. Laws in Switzerland require the IV to provide rehabilitation services, while the SUVA covers the economic consequences of accidents or occupational diseases [72,73]. The complexity and fragmentation of the Swiss social insurance system [74] could hinder the implementation process. This is supported by Schusselé Filliettaz et al [75], who found in a survey on integrated care initiatives in Switzerland between 2015 and 2016 that chronic care integration presents a notable challenge due to issues with the allocation of responsibilities, payment structures for health care professionals and institutions, and financing mechanisms. This is further complicated by the different conditions that apply to the assumption of costs by the various cost bearers within the social insurance system. The assumption of costs by the IV, which was emphasized in the interviews, requires the robotic arm to fulfill the criteria effectiveness, appropriateness, and economic efficiency [76]. In addition, the List of Remedies and Equipment (MiGeL) outlines the device categories eligible for coverage and specifies the maximum reimbursement amounts, including approximately 20 groups of approved devices [77]. Currently, no category that includes robotics as an AT is listed in the MiGeL. Without a designated category, similar ATs listed on the MiGeL may receive compensation, potentially leading to higher costs or inadequate support compared to the robotic arm. According to MacLachlan et al [78], the absence of a clear framework for ATs such as robotics hinders implementation efforts, further compounded by a lack of familiarity and knowledge regarding disability and impairment. The clarification of competence and the assumption of costs among the various cost bearers need to be addressed for AT and the robotic arm in the Swiss context.

Moving forward with the implementation of the robotic arm, a focused approach targeting the identified barriers is necessary. The key recommendations from the authors for addressing the most relevant barriers currently identified are listed below:

Moving forward, an implementation team overseeing and facilitating key activities in the selection and implementation is needed [79]. Representatives of implementation teams should encompass various viewpoints, including those of practitioners, supervisors, administrative leaders, and policy makers. The relevance of implementation teams was also acknowledged by Metz and Bartley [79], stating that in a survey of implementation frameworks, 17 (68%) of 25 frameworks emphasized the formation and use of implementation teams as essential elements of the implementation framework to ensure high-quality implementation. Furthermore, a stakeholder analysis by Reed et al [80] would allow a thorough investigation of relevant stakeholders for overall stakeholder engagement and their change over time [40,81-83] for the implementation of ATs.

In addition to the implementation context, successful evaluation of innovation and implementation is needed. In this research, the combination of innovation and implementation outcomes was given as an implementation outlook, as suggested by Hung et al [84]. According to Albers et al [40], clinical outcomes may be inaccurate or poorly understood if implementation outcomes are not measured, making it crucial to combine both types of assessments. Previous research emphasizes the construct relative advantage of the CFIR as a decisive factor for successful adoption and implementation of the innovation [85]. This suggests that successful implementation is more likely when users recognize a clear advantage in the innovation’s effectiveness. Stakeholders must clearly observe the benefits of the innovation, emphasizing the importance of efforts to demonstrate these benefits [57]. Therefore, giving early attention to the relative advantages in implementation planning could be a strategic approach and should be considered as we move forward with the robotic arm and its implementation.

The use of a hybrid design, combining elements of clinical effectiveness and implementation research to explore not only what works but also how, where, and why, was further proposed in the implementation outlook. This would allow to incorporate complexity to establish external validity and gain insights on scaling up and disseminating [86]. The use of mixed methods designs would enable exploratory and confirmatory research, according to Palinkas et al [87], incorporating the perspectives of potential consumers, practitioners, and clients. To assess innovation effectiveness, standardized ADLs were suggested in this research. Integrating outcomes with the International Classification of Functioning, Disability, and Health could further refine functional evaluation tasks, ensuring a comprehensive assessment that goes beyond basic activities and allows for better comparison across research groups [88]. These considerations for methodology, including design and data collection techniques, need to be incorporated in the future research on the robotic arm’s effectiveness and implementation and comparable ATs.

Some limitations should be considered together with the results of this research. First, although efforts were made to include a broad coverage of stakeholder groups, the distribution of representatives varied, potentially introducing bias with certain perspectives dominating the discussion [42]. To decrease this group dynamic bias, sampling procedures were used to group similar perspectives together in focus groups. However, the absence of a family caregiver representative within the nursing or care stakeholder group and representatives from the political perspective group was acknowledged as a limitation, as this may have led to missed important insights. Second, the study focused on implementation within the Swiss social insurance system, limiting its generalizability in terms of governmental financing of the robotic arm and similar ATs. Furthermore, the direct end users of the robotic arm considered in this study are persons with impaired upper extremity due to tetraplegia. However, the target population could be broadened to include individuals with upper extremity impairments caused by congenital conditions, such as cerebral palsy, as well as other medical conditions. Expanding the target group will require further research to address the specific needs and challenges associated with these conditions.

The analysis of the contextual factors was conducted early in the development of the robotic arm. This is a strength, as it provides valuable insights and identified barriers and facilitators that should be considered moving forward. However, further exploration and development may highlight additional aspects influencing context, and identified contextual factors need to be reassessed with the process. Another strength of this research is its potential applicability to future AT implementation processes in Switzerland and countries with comparable social and health insurance systems. The use of the CFIR framework allows for building on these results and integrating findings from other studies, contributing to knowledge about the interaction between innovations and context. Another strength of this study is the contribution of valuable insights to the implementation of ATs, specifically assistive robots.

This research explored the current contextual factors influencing the implementation of an assistive robotic arm in the everyday lives of individuals with tetraplegia in Switzerland. As an example of AT, the robotic arm offers promising benefits in terms of autonomy and improved participation for users, positively influencing their quality of life. Challenges such as high costs and regulatory complexities regarding cost coverage in its implementation need to be addressed, highlighting the need for a framework for implementing ATs. The need for political lobbying, stakeholder engagement, and user involvement throughout the implementation process were identified to facilitate the widespread adoption of ATs in Switzerland and other countries with comparable social and health insurance systems.