Nursing Staff's Attitudes, Needs, and Preferences for Care Robots in Assisted Living Facilities: A Systematic Literature Review

Care robots have been proposed in response to nursing shortages in assisted living facilities (ALF), in which the population of older adults is growing. Although the use of care robots can improve the health of these older adults, their introduction fundamentally changes the work of nursing staff and has implications for the entire healthcare system. In the development of such gerotechnology, it is important to include end-users, but so far, the perspectives of nursing staff have largely been ignored. We have conducted a systematic review to examine the literature on nursing staff's attitudes, needs, and preferences for care robots in ALFs, to guide future research. Using the PRISMA method, we identified 15 publications. We found that nursing staff desire robots that can assist with physically demanding tasks and reduce workload. Further research is needed on nursing staff's concerns and the contextual factors that influence nursing staff's perspectives of care robots.


INTRODUCTION
It is estimated that the U.S. population of older adults -those aged ≥65 years -will grow to 95 million by 2060 [42].This trend in aging will be associated with increased care demands: 70% of older adults will eventually require some form of long-term care [41].At the same time, long-term care facilities face a severe nurse stafng shortage, exacerbated by the COVID-19 pandemic.The U.S. Bureau of Labor Statistics estimates that an additional 245,600 employees are needed just to return facilities to pre-pandemic levels [2].Nurse burnout has contributed to this shortage and also resulted from it; burnout is associated with increased turnover, high workload, and inadequate stafng [12,22].Assisted living facilities (ALF), or residential settings that provide long-term care to older adults, show higher rates of nurse burnout and turnover than do all other healthcare settings [2,22].
Care robots, which include both social and assistive robots, show promise for addressing the mismatch between the aging population and the shortage of professional caregivers.Social robots, which include companion robots, perform work tasks based on "interactional performances" between humans (e.g., caregivers and patients) or with pets [29, p.258].Assistive robots (also referred to as mobility or service robots) perform physical work tasks, such as lifting patients, helping with activities of daily living, or assisting at mealtimes [29].Some evidence suggests that care robots may improve the general well-being of older adults [7,25,39].In a review of 69 studies, the most commonly reported therapeutic benefts were improved mood and emotional states and increased social interaction [39].
Despite such positive efects, the perspectives of nursing staf on the use of care robots have been largely ignored.In the review of 69 studies, only 15 included nursing staf as participants [39].In those 15 studies, the robots relieved nurses of certain tasks (e.g., responding to call lights [26]), but they often required the assistance of staf for operation (27/69; 39%) and thereby increased workload [30,39,40].A scoping review on the efects of care robots on professional caregivers supports these fndings: the introduction of care robots had both positive and negative efects, reducing physical and mental demands in some scenarios and increasing workload in others [29].In addition, high workloads and negative attitudes have been identifed as barriers to staf's adoption of care robots [18,20].These examples highlight the fact that although care robots have the potential to improve the health of older adults, such benefts are in part due to skilled use by nursing staf [34].Further, the implications of introducing care robots into ALFs extend beyond the recipient of care (i.e., the older adult) to the entire healthcare system.Care robots will fundamentally change the nature of nursing work.These fndings are particularly important, given today's climate of widespread nursing shortages and burnout.
Successful deployment of care robots into healthcare systems and facilities requires a comprehensive understanding of multiple stakeholders' perspectives of their use, and their design must be informed by an awareness of the context in which they will be used [9,32].Studies of human-robot interaction have incorporated participatory and user-centered designs, but they have mainly included the perspectives of receivers of care (older adults) [23,31,37] rather than nursing staf.Nurses' perspectives have not been fully implicated in the design, development, and implementation of care robots.Without consideration of nursing staf and the care environment of ALFs, care robots may further exacerbate the nurse stafng crisis and are unlikely to be adopted into care.
To address these research gaps, we have conducted a systematic literature review in order to answer the following research questions: (1) What is known about the attitudes, needs, and preferences of nursing staf in ALFs in relation to the use of care robots?(2) What research methods, designs, and populations have been used in this research?(3) What are the gaps in the literature that warrant future research?

METHODS
We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 protocol [28].(See Figure 1.) Literature Search.On May 12th, 2023, we searched the databases PubMed, CINAHL Plus with Full Text, PsycINFO, IEEE Xplore digital library, and ACM Digital Library, using the following search terms: ("robot*") AND ("senior living facilit*" OR "residential facilit*" OR "independent living" OR "assisted living" OR "senior living center*" OR "nursing home*" OR "skilled nursing facilit*" OR "intermediate care facilit*") AND ("aged" OR "older" OR "elderly") AND ("nurse*" OR "nursing" OR "staf" OR "professional caregiver*" OR "professional carer*") AND ("perspective*" OR "preference*" OR "need*" OR "user-centered design" OR "user-driven design" OR "participatory design" OR "co-design" OR "usability" OR "universal design" OR "user experience*").We set no limit for years of publication.PubMed, CINAHL Plus with Full Text, and PsycINFO were searched by titles and abstracts, IEEE Xplore digital library was searched by metadata, and ACM Digital Library was searched using the 2012 ACM Computing Classifcation System with the flter "Robotics." This search strategy was adopted from a previous review [39].The fve databases yielded 231 publications (53 from PubMed, 44 from PsycINFO, 42 from CINAHL, 35 from IEEE, and 57 from ACM).When these publications were combined, 11 duplicates were identifed and removed using Rayyan, an electronic screening tool [27]; a total of 220 non-duplicate publications remained.
Literature Screening.The frst author screened the 220 publications by title and abstract using predetermined inclusion and exclusion criteria.To be included, the publications had to meet the following criteria: (1) full text written in English, and (2) focus on the predevelopment phase, gathering information on nursing staf's attitudes, needs, and preferences regarding care robots for ALFs.Following a previous literature review [39], we adopted the National Library of Medicine's Medical Subject Headings defnition of robotics: "the application of electronic, computerized control systems to mechanical devices designed to perform human functions" [24].Smart assistive devices and ambient assisted living technologies without a robotic platform were thus excluded from the review.Studies of the implementation of a care robot or of perspectives of an already developed care robot were also excluded.We defned assisted living facility as any residential setting that provides longterm care to older adults, consistent with prior literature reviews [7].Publications that did not focus on such facilities (e.g., aging in place) were excluded.Publications were also excluded if they were not peer-reviewed empirical studies.A total of 185 publications were excluded at this phase, resulting in 35 for full-text review.The frst author then screened these 35 publications by the same inclusion and exclusion criteria, and 22 were excluded; a total of 13 remained.Two additional publications were added via citation searches of the included studies, for a total of 15 publications [4-6, 8, 10, 11, 13, 14, 16, 17, 19, 21, 33, 35, 36].
Six studies (40%) used a predetermined defnition of "care robot, " meaning that aspects of the robot had been decided before the study.Three of these studies focused respectively on robots for a specifc purpose: feeding [4], showering [19], and personal mobility [14].Four studies involved demonstrations with robotic prototypes [8,14,21,36].
Although assisting with physical tasks was the most desired function of care robots, participants in two studies discussed the care that is lost when such work is delegated to robots [4,19].In HRI '24 Companion, March 11-14, 2024, Boulder, CO, USA one of those studies, nursing staf discussed how bathing patients involves more than just hygiene-staf used this time to connect with patients and establish relationships, to motivate patients to participate in the process, and to monitor health changes (e.g., skin breakdown) [19].In the other study, staf expressed concerns about the robot replacing important bonding during meals [4].
Opinions on whether a care robot could provide social support were mixed.In a few studies, nursing staf believed that care robots were capable of providing such care [16] or a limited version of it [8].In one study, staf had a "wait and see" attitude about robots' abilities [14].Although their thoughts about robots' ability to provide social care were inconclusive, nursing staf agreed that delegating such care to a robot would have serious implications [8,10,14,16,19,21,35].In three studies, staf were adamant that robots should not provide social care [14,19,35].In three other studies, staf were less decisive about whether robots should provide social support and feared their introduction could lead to inhumane or subpar care; however, these participants saw robots as a potential mediator of human social interaction [8,16,21].For example, participants in one study discussed the beneft of using a robot to connect residents with loved ones during the COVID-19 pandemic [16].In one study, beliefs about whether care robots could or should provide social support difered on the basis of the staf's cultural background [10].
Finally, participants desired robots personalized to older adults' abilities, routines, preferences, and needs [4,19,21,33,35,36].To achieve personalized robots, one study suggested "end user programming" where nursing staf (i.e., the domain experts) could easily customize robots for older adults [35].This study suggested that robots should have the ability to learn and adapt from previous experiences [35].Another study suggested that nursing staf's intimate relationships with residents could be used to gain knowledge and personalize robots for specifc older adults [33].

Concerns
Nursing staf members voiced several concerns about adopting care robots in ALFs.Ethically, there was disagreement about who should control the robots-the older adults or the nursing staf.In two studies, staf argued that they should supervise and have ultimate control over the robots, instead of the older adults [21,35].Their primary argument for needing control of the robots was to ensure older adults' safety.In other studies, however, nursing staf were concerned about maintaining the older adults' autonomy and dignity [16,19,36].An emergency shut-of button was suggested as a method to increase autonomy and safety [4,6,8,16,19,33].In two studies, staf were concerned about surveillance and emphasized the importance of protecting the privacy of older adults and staf [16,21].Finally, nursing staf worried that robots could replace human caregivers and take away their jobs [5,10,19].
Accessibility was also a common concern across studies.Caregivers in one of the studies thus advocated for "distributive justice, " meaning that everyone should have access to care robots [19, p.29].Care robots can be extremely expensive, but fnances (whether those of the older adult or of the facility) should not prevent someone from accessing these devices [16,19].A staf member in one study commented that the high cost of care robots can lead to extreme precautions in order to keep the robots in good working order [16].As a result, the robots were often locked up when they were needed most (i.e., when family or staf were unavailable) [16].In addition to being fnancially accessible, robots must be easy to hear and see for those with defcits [6].
Finally, staf were concerned about operability.Nursing staf were concerned that care robots might increase workload, and so they desired robots that would be easy to use [8,14,16], easy to clean and maintain [6,16], fexible [8], and reliable [6].Three studies emphasized that robots should save nursing staf time, which could then be invested into providing relational care [6,16,19].Also related to operability, three studies pointed to a lack of education and training on care technologies in nursing curricula [6,8,21].

Potential Factors Infuencing Individuals' Perspectives of Robots
In fve studies, the authors examined potential factors in participants' perspectives of care robots-culture, education, and the participants' role in the facility.In one study, Finnish staf reported signifcantly greater fear that robots would cause inhuman care and increase loneliness, whereas Japanese staf were more likely to believe that robots could reduce anxiety and loneliness [10].In another study, clinicians (i.e., registered nurses) desired robots to help with medications, safety, and mobility tasks, whereas caregivers (i.e., certifed nursing assistants) desired robots to assist with activities of daily living [17].Medical students in one study were concerned about privacy, whereas nursing students were concerned about robots' social functions [21].In two studies with residents and nursing staf, the residents felt more positively towards the robots than did the staf [5,6].Both the residents and nursing staf desired assistance from care robots: residents desired assistance with managing their health (e.g., medication reminders) and nursing staf desired job assistance (e.g., escorting residents) [5].

DISCUSSION AND CONCLUSION
Several themes emerged from our examination of 15 studies on nursing staf's attitudes, needs, and preferences with regard to the use of robots in ALFs.First, there was a consensus that care robots should serve as nursing assistants to help reduce high workload; they should not replace nursing work.Assistance with physical tasks was the most desired function.Whether robots could or should assist with social tasks remained a question.However, a previous literature review has identifed an overrepresentation of social robots as opposed to assistive robots [39].This discrepancy illustrates the necessity of involving end-users at the early stages of care robots' development.Although older adults are an important end-user group, nursing staf are also end-users of this technology.Indeed, our fndings call for future research on what constitutes nursing work in ALFs.The reviewed studies demonstrate that nursing work is extremely diverse, and much of nursing's most essential care work goes unseen due to gender and power dynamics [1,3].
Many important yet invisible aspects of care cannot be accounted for by the technical afordances of a robot.Although care robots may be useful for relieving nurses of physically demanding and menial care tasks, they cannot replace nurses' interpersonal labor.The same is true of the electronic healthcare record, which has been widely adopted throughout the healthcare system but reduces nurses' face-to-face interactions [15].To develop ethical, efective, useful care robots, it is important for researchers to gain a deep understanding of nursing work and the benefts and consequences associated with introducing a robot into the setting of an ALF.Second, nursing staf were concerned about the ethics, safety, accessibility, and operability of care robots in ALFs.Opinions about who should have control over the robots were mixed.On the one hand, giving older adults control over a care robot may enhance autonomy, but, on the other, nursing staf are responsible for ensuring residents' safety.Further research on how to preserve older adults' autonomy while balancing it with safety is needed.Robots' operability was especially important to nursing staf, who did not want robots to increase their already high workloads.Future research should focus on understanding nurse workfows and how to thoughtfully implement robots in a way that does not increase workload.In addition, in the context of the COVID-19 pandemic and widespread stafng shortages, where nurses already feel undervalued, it is important robots are introduced as assistants to nurses, rather than replacing nursing work.
Third, there were methodological limitations identifed across the 15 studies.The sample sizes were small; 53% of the studies had fewer than 10 participants.Several diferent questionnaires were used to assess nursing staf's attitudes, needs, and preferences, making it difcult to compare fndings across studies, and although some of these instruments have been validated, many have not.Further research is needed to develop and validate relevant instruments for understanding nursing staf's perspectives of care robots.Six studies used predetermined defnitions of care robots, which limited the implications of their participants' comments.To fully understand the perspectives of nursing staf at the early stages of robot development, one might need nurses to describe their ideal robot without any preconceived ideas.Also, the studies provided very limited demographic information on the nursing staf or facilities.Only fve studies considered demographic factors when interpreting their results, and fndings difered across culture, education, and job title.To understand the factors that infuence attitudes toward and needs or preferences for the use of care robots, it is important to collect and consider demographic information on nursing staf as well as other end-user groups.This can promote the development of care robots that are personalized and useful, as well as increase the likelihood that care robots will be adopted into practice.Six of the studies included multiple facilities, but with little exploration of how diferences among facilities might infuence users' perspectives of care robots.Studying a single facility may allow detailed exploration, but the inclusion of multiple facilities allows comparisons, with greater potential for generalization.
In this systematic review of 15 studies on nursing staf's attitudes, needs, and preferences with respect to the use of care robots in ALFs, we have found that nursing staf desire care robots that will assist with physically demanding tasks and reduce workload.But the nurses wanted to be able to preserve their interpersonal caring work too.Consideration of ethics, safety, accessibility, and operability informed nurses' concerns across the studies.Further research is needed to mitigate these concerns and understand the demographic, social, environmental, and organizational factors that infuence individuals' perspectives of care robots and their adoption in ALFs.