This systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO; registration CRD420251079714). This review was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines [17]. All processes were independently evaluated by 2 reviewers, and any disagreements were resolved by discussion with the corresponding author.

The inclusion criteria for the studies were as follows: studies must include HCPs as participants; examine all related perspectives, such as attitudes, perceptions, acceptance, views, or willingness to use SARs; use qualitative, quantitative, or mixed methods research designs; and include original empirical studies involving human participants. Studies were excluded if they focused on non-HCP populations, did not primarily investigate SARs, or were not original research. The eligibility criteria for the inclusion and exclusion of studies are detailed in Textbox 1.

A comprehensive literature search was conducted across 13 databases, with the final search completed on July 29, 2025. The following databases were systematically searched: PubMed, Cochrane Library, Scopus, IEEE Xplore, ScienceDirect, CINAHL with Full Text (EBSCO), Epistemonikos, MEDLINE (OVID), Web of Science, Embase, UpToDate, and CEPS. In addition, ProQuest Dissertations and Theses Global were reviewed to capture relevant gray literature.

Search strategies were formulated using the Population, Intervention, Comparison, and Outcome (PICO) framework and were tailored for each database. A combination of MeSH (Medical Subject Headings) and free-text terms was used along with Boolean operators. The keywords incorporated “Healthcare Professionals,” “Socially Assistive Robot,” “Acceptance,” “Perspective,” “Attitude,” “Perception,” “Views,” and “Healthcare Setting.” Overall, 2 authors independently reviewed and validated the search process to ensure its accuracy and comprehensiveness. The complete search strategy for all databases is provided in Table S1 in Multimedia Appendix 1.

The 2 reviewers independently performed the identification, screening, eligibility assessment, and inclusion stages. The records were sourced from database searches and supplementary sources, and duplicates were removed using EndNote (version 21; Clarivate). Titles and abstracts were screened to exclude irrelevant studies, and the full texts of potentially eligible articles were subsequently evaluated. Reasons for exclusion at the full-text stage were documented, and any disagreements were resolved through consultation with the corresponding author to reach a final decision. Studies that satisfied all inclusion criteria were included in the final synthesis. The overall study selection process is depicted in the PRISMA flow diagram.

Data extraction was independently conducted by 2 reviewers. Disagreements were resolved through discussion with the corresponding author to reach a consensus. In instances in which clarification was necessary, the authors of the original studies were contacted. The extracted variables included study details (author, year, and country), participant information (professional type, workplace, and number of HCPs), study purpose, methodology, and reported outcomes related to acceptance, attitudes, perceptions, or willingness to use SARs.

The quality of all included studies was evaluated using the Mixed Methods Appraisal Tool (MMAT) [18]. This process involved a systematic assessment of the risk of bias across 5 domains: selection bias, measurement bias, response bias, confounding bias, and reporting bias, with all evaluations thoroughly documented. The 2 reviewers independently conducted the appraisals, and any discrepancies were resolved through consultation with the corresponding author to achieve consensus. The methodological quality of each study was quantified using a scoring system ranging from 0 to 5 points [16], with criteria rated as “Yes” (1 point), “Unclear” (0.5 points), or “No” (0 points). Studies identified as having a moderate risk of bias were penalized by 1 point, whereas those with high risk of bias were penalized by 2 points. Based on the final scores, studies were classified into 3 levels of quality: high quality (4.5-5 points), moderate quality (3-4 points), and low quality (0-2.5 points).

During synthesis, study quality was explicitly considered; findings from high-quality studies were assigned greater interpretive weight, whereas results from low-quality studies were reported with caution and were acknowledged as less reliable. Potential confounders, including demographic and contextual variables, were identified and discussed to avoid overgeneralization. A narrative synthesis approach was used to integrate heterogeneous study designs while simultaneously recognizing methodological constraints. Transparency was ensured by reporting the distribution of the study quality (low, moderate, and high) and highlighting areas where insufficient rigor may have influenced the evidence. Collectively, these procedures minimized the influence of bias, strengthened validity, and supported a balanced and cautious interpretation of the findings.

Qualitative and quantitative data were analyzed independently before integration. For the qualitative strand, data extracted from the included studies were synthesized using thematic analysis. The material was systematically extracted, inductively coded, and organized into categories, with themes refined iteratively through discussion until consensus was achieved. Quantitative findings were aggregated and summarized descriptively. Integration of the 2 strands followed a convergent parallel approach, whereby qualitative and quantitative results were analyzed separately but compared concurrently through narrative integration. This allowed for the identification of areas of convergence, divergence, and contradiction across data types. The integrated findings are presented in the Synthesis section and subsequently mapped onto the UTAUT theoretical framework [19] according to its respective dimensions.

A total of 516 studies were retrieved from 11 electronic databases, including PubMed, Cochrane Library, Scopus, IEEE Xplore, ScienceDirect, CINAHL, Epistemonikos Database, MEDLINE, Web of Science, Embase, and UpToDate, with an additional 9 studies identified through manual searching as of July 29, 2025. Following the removal of 67 duplicate entries, 458 records were screened by 2 independent reviewers. Of these, 426 were excluded based on the title and abstract that did not meet the eligibility criteria. Full-text versions of 32 articles were sought, of which 2 articles were unavailable for retrieval. The remaining 30 full-text articles were evaluated against the inclusion and exclusion criteria, resulting in the exclusion of 15 studies. Ultimately, 15 studies were included: 6 qualitative, 5 quantitative, and 4 mixed methods studies (Figure 1). The reasons for exclusion included the type of study (opinion article), lack of examination of SARs, absence of focus on HCPs, or failure to address relevant outcomes (Table S2 in Multimedia Appendix 1).

This systematic review encompassed 15 studies conducted in diverse countries, including Israel, England, Colombia, Taiwan, Egypt, Canada, South Korea, Australia, Slovenia, Singapore, and France. These studies investigated HCPs’ perceptions, acceptance, and attitudes toward the implementation of SARs in health care environments. Collectively, the study involved 3166 HCPs with individual study sample sizes ranging from 10 to 1284 participants. HCPs represented a broad spectrum of professions, with the majority being nurses and midwives (1960/3166, 61.9%). The primary settings for these professionals were hospitals (1688/3166, 53.3%) and primary care facilities and nursing homes (845/3166, 26.7%). Most participants were female (2618/3166, 82.7%) and originated from Europe (1709/3166, 54%) and Asia (1266/3166, 40%; Table 1). Furthermore, among the 15 studies included in this review, 8 studies involved participants with no prior direct experience with SARs [20-27]. In 2 studies, participants were shown videos related to SARs before engaging in research activities [28,29]. A total of 5 studies involved direct experience with the actual use of SARs [30-33].

In reporting the results, we explicitly integrated both qualitative and quantitative evidence under the UTAUT framework. To enhance transparency and rigor, the findings were contextualized according to the methodological quality of the contributing studies, as assessed by the MMAT. Subgroup differences (eg, professional role, setting, region, sex, or educational background) were reported, and these were highlighted to provide more targeted insights. Throughout the synthesis, conclusions from higher-quality studies are presented with greater evidentiary weight, whereas results from lower-quality studies are interpreted with appropriate caution.

A total of 12 studies were rated as moderate quality [20,22-32] in Table 1, one study as high quality [21], and 2 studies as low quality [33,34]. A detailed description of the quality appraisal process, including the assessment criteria and individual study ratings, is provided in Table S3 in Multimedia Appendix 1.

The primary sources of bias across the included studies pertained to sampling strategies, representativeness, measurement validity, transparency, and the comprehensiveness of reporting. A high-quality study was distinguished by a large sample size and the use of validated measurement instruments. In contrast, studies rated as low quality were characterized by small sample sizes, reliance on self-developed questionnaires without evidence of validity, and reporting bias (Table S4 in Multimedia Appendix 1).

Most of the studies included in the analysis used convenience sampling, often using relatively small samples sourced from a single institution, which may limit the generalizability of their results. However, 1 cross-national study examined whether cultural differences across countries affected HCPs’ training needs for SARs by gathering data from HCPs in 18 countries, with each country contributing between 40 and 140 participants, thereby facilitating cross-cultural comparison [26]. Measurement bias was also observed, as several studies used self-developed instruments without clear reporting of their reliability or validity [26,27,33].

Concerns regarding response bias were identified, as nonresponse rates were rarely reported; only 1 study explicitly documented a nonresponse rate of 25% [25]. Potential confounding factors, such as demographic variables, were infrequently controlled for in quantitative studies, and all qualitative studies relied on subjective interpretations. Some studies with methodological limitations nonetheless reported statistically significant findings [33,34], indicating that such results should be considered within the context of their study designs, and interpretations should be made cautiously.

This systematic review investigates HCPs’ perspectives on the clinical application of SARs. This study synthesizes key issues, including the challenges and barriers faced by HCPs in using SARs, their levels of acceptance, influencing factors, ethical literacy, educational needs, and considerations for clinical implementation. Additionally, this review offers recommendations for the implementation of SARs.

This review highlights the diverse challenges faced by HCPs when integrating SARs into clinical practice, with technological reliability standing out as a central concern. Commonly reported issues include hardware malfunctions, software errors, limited battery capacity, and unstable connectivity. Such problems can interrupt clinical workflows and undermine HCPs’ trust in SARs. These findings echo Masala and Giorgi [35], who demonstrated a strong link between perceived reliability and trust.

Beyond technical limitations, concerns also extend to SARs’ ability to interpret emotional expressions and nonverbal cues. Many HCPs perceive these systems as lacking empathy and emotional intelligence qualities essential for person-centered care. While acknowledging that SARs can provide valuable support for routine tasks, HCPs remain doubtful about their ability to preserve the interpersonal aspects of caregiving, a concern similarly raised by Zuschnegg et al [36]. Respondents also questioned whether SARs could demonstrate sound judgment and adaptability in complex clinical situations.

Professional identity represents another source of tension. Some nursing staff expressed concern that robots might intrude on core nursing responsibilities, potentially threatening their professional roles. This ambivalence reveals a paradox; on one hand, HCPs emphasize the irreplaceable value of human caregiving, and on the other, they anticipate institutional pressures to replace human labor with robotic alternatives. Importantly, such perceptions may be shaped less by direct experience with SARs and more by media portrayals or imagination.

Overall, it suggests that HCPs tend to view SARs as complementary tools rather than substitutes, consistent with the findings of Madi et al [37]. However, research results presented a counterpoint, arguing that SARs may fail to reduce workload and could even add to it, due to the resources needed for implementation and maintenance [38]. This raises concerns about their long-term sustainability. Nonetheless, in the context of workforce shortages, SARs could still play a meaningful role by alleviating routine burdens, provided that future designs emphasize intuitive human-robot interaction (HRI) and user-friendly functionality. Such improvements would ensure that SARs are seen as practical assistants that enhance, rather than hinder, clinical practice.

This review also found that the role of education in shaping HCPs’ acceptance of SARs varied across studies. These inconsistencies may stem from differences in study quality, regional context, or the professional backgrounds of participants. Interestingly, age did not appear to have a significant influence, which contrasts with the UTAUT, where age is considered a key moderating factor. This discrepancy may be explained by HCPs’ prior experience with technology. This conclusion is supported by the research findings of Naneva and colleagues [39], who determined that familiarity with similar technologies enhances the willingness to adopt SARs.

Although education level and age are often viewed as important moderators in UTAUT, current evidence on their impact remains inconclusive. Future research should use more rigorous methods to clarify the role of these variables in shaping HCPs’ acceptance of SARs.

The reviewed studies indicate that HCPs express concerns regarding “blurred boundaries” and “compromised patient dignity” in HRI, although empirical evidence in this domain remains limited. The specific challenges identified include the absence of adequate regulations for AI-driven decision-making, risks associated with algorithmic bias, and issues pertaining to data privacy. For instance, under the General Data Protection Regulation, cross-border transmission of patient data necessitates explicit consent and encryption. Furthermore, if the training data are not representative, algorithmic bias may ensue, leading to AI recommendations that disadvantage specific groups [38]. These concerns are consistent with broader calls for robust ethical guidelines and governance frameworks to ensure the “responsible adoption” of AI in clinical settings [40-42].

Building on these perspectives, this review underscores the necessity for ethical frameworks governing SARs to integrate the principles of transparency, accountability, and inclusivity throughout the design, development, and clinical deployment of AI systems. The incorporation of ethical literacy serves as both a professional competency for HCPs and an institutional safeguard, thereby mitigating potential risks and fostering responsible innovation. This argument aligns with the conclusions of El Arab et al [43] and Zhang et al [44]. Practical strategies may involve the establishment of “robot ethics committees” by convening HCPs, relevant experts in ethics and law, and technology developers within hospitals to oversee data use, informed consent processes, and cultural sensitivity in HRI. At the policy level, governments might consider forming national “AI ethics committees” to regulate technological trajectories and amend existing laws as necessary, as also proposed by Zhang et al [44].

Notably, Mlakar et al [25] identified a positive correlation between ethical acceptability and overall acceptance of SARs by HCPs, highlighting the potential influence of ethics on adoption decisions. Collectively, these findings indicated that ethical literacy directly affects clinical integration. Consequently, future research should aim to validate the extent to which enhancing HCPs’ ethical literacy facilitates the adoption and sustainable use of SARs.

Beyond individual attitudes, organizational support for education and resource investment is crucial in influencing HCPs’ acceptance and adoption of SARs (Figure 3). The reviewed studies consistently demonstrate that successful clinical integration necessitates comprehensive training programs that encompass operational skills, AI literacy, and ethical literacy. Among these, proficiency in daily operations, such as process flow, troubleshooting, and maintenance, has emerged as the most immediate requirement. A deficiency in these competencies may lead to clinical setbacks and diminished willingness to engage. This finding aligns with that of Langer et al [45], who emphasized that understanding both the functions and limitations of SARs is a key prerequisite for effective implementation.

This review emphasizes the necessity of integrating ethical literacy as a core component of SARs training, addressing patient autonomy, data privacy, informed consent, and the equilibrium between efficiency and empathy in care. Enhancing these competencies can assist HCPs in forming realistic expectations and reducing their frustration in clinical applications. The study further contended that as AI and robotics increasingly influence patient safety, quality of care, ethical decision-making, data privacy, and resource efficiency, training requirements must extend beyond basic technological literacy [46]. Therefore, it is recommended to incorporate AI and robotics literacy into training curricula, starting with undergraduate and prelicensure education. The research endorsed this approach, asserting that such measures are essential for future HCPs to effectively interpret AI outputs and apply them in practice [43].

Despite these varied needs, most institutions currently lack formal training programs dedicated to SARs, with learning often occurring informally through peer-to-peer exchange. This reliance risks neglecting essential competencies in the technical, ethical, and HRI domains. Addressing these gaps requires the development of flexible, interdisciplinary training initiatives or guiding principles that can be adapted to local contexts. Such initiatives would enhance professional preparedness for SARs integration while remaining aligned with the overarching goal of delivering safe, efficient, and person-centered care [43].

Training requirements differ significantly across professional roles, highlighting the need for customized approaches. Nursing professionals often prioritize practical operational competencies such as process flow, troubleshooting, and patient-facing interaction skills. Conversely, managers tend to focus on workflow integration, workload distribution, and resource allocation. Meanwhile, physicians and rehabilitation specialists should concentrate on clinical applicability, patient safety, and interdisciplinary collaboration. Consequently, training programs should be modular and tailored to specific roles, ensuring that each professional group acquires the competencies most relevant to its responsibilities. This reliance may result in the neglect of competencies in the technical, ethical, and HRI domains, highlighting the urgent need for standardized, interdisciplinary training initiatives supported by institutional policies and practice guidelines [43].

It is imperative to enhance the early integration of technological literacy into health care education. AI has the potential to revolutionize professional learning and augment digital health competencies among prospective clinical staff members. Therefore, structured curricula should emphasize foundational AI concepts, ethical principles, and simulation-based training. Educators should also be trained to ensure the effective incorporation of AI into both clinical and educational settings [43].

The triangular framework illustrates the interrelated educational needs of managers, HCPs, and undergraduates. These include operational skills, AI literacy, and ethical literacy, which are essential for the responsible integration of AI and robotics into health care practice.

The successful integration of SARs into clinical practice necessitates consideration of both environmental preparedness and organizational support. Infrastructure readiness is essential, and clinical environments must be modified to accommodate SARs through suitable physical layouts, charging facilities, and docking stations. Engaging HCPs in the co-design process can enhance feasibility, promote ownership, and increase acceptance. Implementation may also benefit from phased incremental strategies, starting with pilot projects, and providing continuous feedback to ensure sustainability and adaptability. A clear delineation of roles and responsibilities is crucial, specifying which tasks SARs should perform and which remain within the purview of HCPs. This aligns with Elendu et al [47], who emphasized transparent decision-making, clear attribution of responsibility, and structured task allocation.

In addition to workflow integration, economic feasibility and policy endorsements are critical determinants. Conducting cost-effectiveness analyses is essential to assess procurement, training, maintenance, and infrastructure in relation to anticipated benefits such as improved efficiency, reduced workload, and enhanced patient outcomes. These considerations align with the findings of Gkiolnta et al [38], who emphasize that although HCPs acknowledge the potential value of SARs, their adoption ultimately hinges on practical viability and institutional commitment.

This review highlights several deficiencies in the extant literature concerning HCPs’ perspectives on SARs. Primarily, existing studies predominantly emphasize short-term outcomes, with insufficient exploration of long-term experiences such as user satisfaction and sustained engagement. Given that many HCPs lack direct interactions with SARs, research frequently depends on video demonstrations or hypothetical scenarios. Consequently, future research should prioritize field trials within authentic clinical settings to produce evidence that more accurately reflects real-world practice.

Second, the propensity to adopt SARs is anticipated to differ among various professional roles and clinical environments, underscoring the need for comparative studies of distinct groups of HCPs. Further research is imperative to evaluate the technical reliability and adaptability of SARs in the medical setting. Interdisciplinary collaboration is crucial, and joint training initiatives between HCPs and robotics engineers may facilitate the alignment of technological designs with clinical requirements. Ethical considerations constitute another significant area of research, necessitating the development of concrete ethical assessment tools and practical guidelines to guide value-driven decision-making.

Interdisciplinary collaboration is essential for the effective integration of technologies, as evidenced by this review that advocates the establishment of joint training programs between HCPs and robotics engineers. Such initiatives aim to foster mutual understanding of clinical needs and technological constraints, thereby enhancing the feasibility of applications. Furthermore, ethical considerations require further exploration, particularly in the development of practical assessment tools and context-sensitive guidelines to support value-driven clinical decision-making.

Hence, cross-national variation is evident in how SARs are perceived and adopted, reflecting differences in health care systems, professional norms, and cultural attitudes toward technology [13]. These factors influence not only levels of acceptance but also the type and intensity of training required. Accordingly, while this synthesis consolidates global evidence, its generalizability remains constrained, and future research should prioritize culturally sensitive approaches to inform policy and practice.

Further research should investigate the impact of “robot literacy” on HCPs’ acceptance and implementation outcomes. At the educational level, previous literature has suggested that medical and nursing schools could explore integrating courses on “AI and robotics applications,” ideally in collaboration with AI and engineering faculties, to help upcoming HCPs develop essential technological competencies. Rather than prescribing rigid, standardized curricula, such training has been proposed in the form of adaptable modules or guiding principles tailored to various professional roles, clinical contexts, and cultural environments. Exploring these approaches in future studies may help address existing knowledge gaps and foster collaboration among HCPs, data scientists, ethicists, and engineers, thereby supporting the co-design of ethical, person-centered, and clinically relevant AI systems [43].

Although UTAUT offered a strong analytical structure, several themes in this review—ethical considerations, professional identity, and organizational readiness—extend beyond its original constructs. These findings suggest that UTAUT may be insufficient on its own to fully capture the complexities of SARs adoption in health care. Future research should consider adapting UTAUT or developing an expanded model that incorporates these dimensions, drawing on complementary theories from implementation science and ethics. By identifying these extensions, this review not only applies UTAUT but also contributes to refining its applicability in health care contexts.

This review had several strengths. By systematically synthesizing evidence from the perspective of HCPs, this study underscores an essential yet frequently underexplored dimension of SARs—the factors influencing HCPs’ acceptance and adoption. This focus is particularly significant as HCPs’ engagement is a critical determinant of whether SARs can be effectively integrated into clinical practice. This review provides a comprehensive account of HCPs’ perceptions, expectations, and concerns by integrating qualitative, quantitative, and mixed methods studies across diverse professional backgrounds and clinical contexts.

The application of the UTAUT as an analytical framework further enhances the interpretive coherence and theoretical grounding of the findings. Collectively, these features augment the relevance and forward-looking value of this review, providing insights that can guide both technological design and clinical implementation strategies.

Several limitations should be noted. A limited number of multimethod studies failed to clearly articulate their methodologies, and qualitative and quantitative findings were reported separately, thereby hindering their integration. In addition, qualitative studies often rely on small sample sizes, which may limit representativeness, while numerous quantitative studies use questionnaires lacking fully validated reliability or validity, potentially compromising explanatory power. In mixed methods research, relatively few studies have concentrated exclusively on HCPs; in several instances, quantitative data were gathered from patients or family members, with HCPs’ perspectives primarily represented through qualitative evidence.

The heterogeneity of study designs and outcomes not only complicated subgroup analyses but also precluded the possibility of conducting a meta-analysis, thereby limiting the strength of quantitative synthesis. Additionally, some studies included only nursing staff, whereas others did not provide detailed information on professional backgrounds, complicating comparisons across roles or disciplines.

Moreover, inconsistencies in quantitative instruments and a lack of comparable data render meta-analysis impractical. Collectively, these limitations underscore areas in which future research could enhance the quality and comparability of evidence. Nonetheless, by consolidating diverse findings and emphasizing the pivotal role of HCPs, this review makes a timely and significant contribution to the understanding of how SARs may be effectively integrated into health care, offering forward-looking insights for research, policy, and practice.

One key limitation is that many included studies involved HCPs without direct experience using SARs. This restricts ecological validity, as the findings largely represent anticipated perceptions rather than insights grounded in real-world interaction.

Another limitation is the quality of the included studies. Among the 15 studies synthesized, only 1 was rated as high quality [21], 2 were low quality [33,34], and the remainder as moderate quality. As a result, our interpretations were made cautiously to avoid overestimating the impact of SARs.

Despite these limitations, this review underscores the centrality of HCPs’ perspectives in shaping the future of SARs and informs the development of policies, training initiatives, and design strategies that can promote responsible and sustainable integration.

This review highlights the pivotal role of HCPs in shaping the effective integration of SARs into clinical practice. Across diverse contexts, HCPs acknowledged the potential of SARs to reduce workloads, improve efficiency, and support patient well-being, while also raising concerns about technological reliability, ethical challenges, and professional boundaries. Ethical literacy and targeted training programs have been discussed as potential enablers of acceptance, underscoring the importance of comprehensive, interdisciplinary approaches that incorporate transparency, accountability, and person-centered care principles.

In addition, methodological inconsistencies were evident, and most of the included studies were of only moderate quality, with limited high-quality evidence. Small sample sizes and the use of nonvalidated instruments further reduce generalizability. Consequently, our interpretations were intentionally cautious to avoid overstating the potential impact of SARs, and the actual readiness of HCPs for adoption remains uncertain. The successful clinical implementation of SARs also depends on broader contextual factors such as organizational readiness, economic feasibility, and clear task delineation to ensure sustainability and trust. While these considerations extend beyond the direct findings of this review, they align with themes highlighted in previous literature and represent important directions for further inquiry. Despite the limitations of the current evidence base, this review contributes a theoretically grounded synthesis using the UTAUT framework and consolidates diverse perspectives of HCPs. Future research should prioritize longitudinal and context-specific evaluations, engage participants with direct SARs experience, and foster interdisciplinary collaboration to generate stronger evidence for the responsible, equitable, and sustainable adoption of SARs in health care.