Abstract
Background and Objectives:
Use of companion robot pets to reduce social isolation and loneliness in older people is well-established. Outcomes associated with the pets are promising in residential care settings but less clear in home settings and among people who do not meet criteria for dementia. Also, low-cost devices, such as the Joy for All pets, may not be as effective.
Research Design and Methods:
Caregiver-older adult pairs (n = 50) in PA’s Caregiver Support Program were recruited for a single-blind pilot randomized controlled trial, in which half received the Joy for All pet immediately and half at the end of the 4-week trial. We used a repeated-measures mixed model to compare differences in caregiver reports of the frequency of positive and negative affect (adapted PROMIS scales).
Results:
The groups were mostly balanced in sociodemographic features, caregiving involvement, and health. Half of the older adults in each arm had a dementia diagnosis. The gain in positive affect was 7.28 points (95% CI, 3.1–11.4) in the intervention arm (z = 3.43, p = 0.001) in a model adjusting for caregiver age, ADL caregiving, and dementia severity (model X2 = 39.1, df = 6, p <0.0001). Dementia severity was associated with lower positive affect. The groups did not differ in negative affect. Greater engagement with pets was associated with higher positive affect.
Discussion and Implications:
A low-cost, robot pet can improve positive affect in older people receiving family care. Caregivers reported that the robot pets were a topic of conversation and focus of social interaction with visitors, which in turn may improve mood. Thus, the pets should be considered an environmental stimulus as much as an interaction partner.
Keywords: Robot, pet, caregiving, dementia, affect
INTRODUCTION
Use of companion robot pets to reduce social isolation and loneliness in older people is well-established. As early as 2001 pet-like social robots were deployed in skilled nursing settings.1 Perhaps the best known is PARO, now in its 8th generation and in use in Japan and Europe since 2003 (http://www.parorobots.com/). Engineering workshops on robot and human interaction have examined design principles for the pets extensively. For example, a 2004 effort examined the need for social robots to mimic human-human dialogue in human-robot dialogue. “Social robots should recognize, respond to, and employ where possible all modalities that humans naturally use to communicate. These include verbal cues such as speech, intonation, and tone of voice, and nonverbal cues such as gesture, posture, and stance, among others.”2
Robot pets must also follow expected communication and behavioral norms. For example, sensors in the head and cheek of the Joy for All robot pet respond to touch and produce a “nuzzling” effect.3 The pet barks or meows when it is touched and its sensor detects a change in light. These “touch- and light–activated sensors … enable autonomous responses through vocalizations and movements for the purpose of social interaction” (Joy for All, 2018. URL: https://joyforall.com/products/companion-cats).4 The Joy for All pets (cats, dogs, birds) are low tech and low cost (~$130) compared to PARO (~$6,000),
Prior research involving robot companion pets suggests that older people prefer realistic animals.5 The pets increase social interaction between participants and care providers, family members, and visitors.6 Older people with higher levels of depression and greater medical fragility appear to be more receptive to having the pets in their homes.3 Older people find the robot pets acceptable and interact with them.7
Outcomes associated with exposure to the pets are promising. A randomized trial among people admitted to a community hospital found lower rates of delirium and falls in a group receiving a robot pet during the hospital stay.8 In a meta-analysis of 7 randomized controlled trials conducted in long-term care sites, depression and anxiety were reduced in the pet arm.9 An RCT using PARO among people with dementia in a residential setting reported benefit in emotional expression compared to a reading group.10 Other RCTs involving PARO have shown benefit in physiologic indicators, such as pulse oximetry and pulse rate,11 and cost-effectiveness.12
However, results are less consistent outside the residential care setting and in people who do not meet criteria for dementia, and also in studies involving low-tech devices, such as the Joy for All pets. Few studies are available. In one single-arm study involving community-dwelling older adults with depression, robot exposure resulted in a decline in depressed mood and anxiety,13 but a randomized controlled trial of community-dwelling older adults with dementia did not show benefit in neuropsychiatric symptoms or agitation.14 Studies reported on the Joy for All website all involve residential care populations (https://agelessinnovation.com/blog/). Other studies do not involve randomized designs, only assess uptake rather than outcomes, or are anecdotal.
For this reason, we conducted a pilot randomized controlled trial of the Joy for All pet in a community setting. We worked with a community Area Agency on Aging and its federally funded Caregiver Support Program to recruit family caregivers and older people. We sought to determine (1) if older people in the community, some with no or only mild dementia, interacted with the pets; (2) if exposure to the pet was associated with improvement in positive affect and reduction in negative affect as reported by family caregivers; and (3) if levels of engagement with pets were associated with changes in affect.
METHODS
Sample
Pairs of family caregivers and older adult care receivers were recruited from the Caregiver Support Program administered by the Area Agency on Aging, Allegheny County Department of Human Services (PA) through a data sharing agreement. Staff from the AAA Caregiver Support Program mailed a letter to participants informing them of the study, with an option to opt out by calling the AAA or contacting the principal investigator. If potential participants did not opt out, the research team contacted them, explained the study, and attempted to obtain informed consent.
Caregiver inclusion criteria included participation in the Caregiver Support Program, aged 18+, care for an older person aged 60+ for at least 6 months, and willingness to receive the robot pet. Family caregivers did not need to live with the older person, and having a pet in the home was not an exclusion. Exclusion criteria included not being able to report back study features in a screening interview. Verbal consent was obtained in a telephone screening interview. The University of Pittsburgh Human Research Protections Office approved the research, which was registered at clinicaltrials.gov.
Care receivers were also consented, as required by our IRB, since the study asked caregivers to report on the older person’s interaction with the robot companions and mood. Inclusion criteria for care receivers included community residence for the duration of the study. They were not required to meet criteria for cognitive impairment. During telephone screening interviews with caregivers, we asked to speak with older adults to explain the study and obtain oral consent. In cases where the older person was not able to come to the phone, caregivers were asked to explain the study and determine willingness to participate.
The Pennsylvania Caregiver Support Program provides financial reimbursement for caregiving related services and supplies. While Pennsylvania’s Program began in 1990, it is currently part of the National Family Caregiver Support Program (NFCSP) and funded by federal Older Americans Act and state funds (via the Aging Block Grant) to Pennsylvania’s s network of Area Agencies on Aging. Caregivers in the program have access to assistance, training, and respite but also reimbursement for ongoing services and supplies related to care.
Intervention and Design
The Caregiver Support Program provided the Joy for All pet free to all participants, who were encouraged to keep the pet after the study period. Participants could choose between adult cat and dog, or kitten and puppy. The pets are instrumented stuffed animals that bark, purr, and move (eyes, head) in response to touch. As the Joy for All website states, “at your touch, they’ll wag, and snuggle, letting you feel their soothing heartbeat” (https://joyforall.com). If ignored, they go into a sleep position. They run on a simple battery. They have no memory, do not ambulate, and do not record aspects of interaction.
Participants were randomized to the immediate (intervention) or delayed (control) group after the baseline interview using a random number generator. A follow-up interview was conducted approximately 4 weeks later with both groups. After the follow-up interview, participants randomized to the delayed access group received the pet. Interviewers were masked to treatment assignment and received training in study measures and responsible conduct of research. Research data were collected exclusively through telephone contact.
Measures
Primary outcomes were positive and negative affect in the older person as reported by caregivers. For the affect measures we used 2 PROMIS scales, Positive Affect – Short Form 15a and Emotional Distress-Anger – Short Form 5a (https://www.healthmeasures.net/explore-measurement-systems/promis). These were adapted for proxy response. For positive affect, we elicited the frequency of 15 emotions (cheerful, attentive, delighted about something, happy, joyful, enthusiastic, determined/focused, interested in what is going on, creative, happy with themselves, peaceful, good-natured, useful, understood, content). Responses ranged from 1, not at all, to 5, very much. For negative affect we asked about the frequency of 5 emotions (irritated, angry, ready to explode, grouchy, annoyed). Responses ranged from 1, never, to 5, always. Questions elicited the frequency of the emotion in the past 7 days in “the person I am providing care for.”
Other measures assessed sociodemographic features, health and disability of the caregiver, features of caregiving involvement and intensity, and dementia severity in the care receiver. Dementia severity was assessed using the Quick Dementia Rating System-Caregiver (QDRS).15 The QDRS follows the format of the Clinical Dementia Rating (CDR). It assess ten categories of cognitive and functional ability, each with 5 options (ranging from unable to fully competent). Using the informant version, caregivers were asked to rate older persons based on a comparison of current abilities to how they used to be. The sum of QDRS scores is used to establish dementia severity, ranging from normal to severe dementia.
We also assessed engagement with the robot pet in the intervention arm. The 6 items were drawn from the Companion Animal Attachment Scale16 and asked caregivers how often the older person did any of the following: “hold, stroke, or pet the animal,” “sleep near or in a room with the animal,” “interact with the animal (talk to it, include it in an activity),” “have a close relationship with the animal,” treat it like a real pet,” and “feel responsible for it.” Answers were summed.
Caregivers in the immediate access arm were also offered an opportunity to describe in their words how older adults interacted with the pets, which we captured in the follow-up interview. We asked, “Any thoughts about how the person interacted with the robot companion pet?” Responses were recorded verbatim.
Caregivers were not instructed to observe the older person in any particular way, and we did not specify or require interaction with the pet as part of the study.
Analyses
Measures were first assessed descriptively, with comparison between study arms at baseline to assess balance. For testing the effect of the intervention, we developed 2 repeated-measures mixed models, one for positive affect and one for negative affect, with adjustment for severity of dementia, whether caregivers provided ADL care (to assess level of involvement by care receivers), and age of caregiver. Data were formatted in long format using STATA with an indicator for group and timepoint. The group by time interaction term indicates the effect of the intervention on outcomes. Given the small sample, we also examined the proportion of participants at follow-up with high scores on the reported affect measures, which we defined as the top quartile in the full distribution of participants.
Given the 2 primary outcomes, caregiver-reported positive and negative affect, we adjusted significance levels using a Bonferroni correction for an alpha of 0.025.
RESULTS
Participation
Participants were recruited between March and October 2024. All were drawn from the Allegheny County, PA Caregiver Support Program. Less than 2% of potential participants opted out in response to the initial letter from the Allegheny County Department of Human Services. Of a total of 294 contacted, 190 (64.6%) declined to participate (either caregiver or older adult), 38 (12.9%) were not reached, and 13 (4.4%) were ineligible. Ineligibility was mainly a result of the older person no longer residing in the community (no longer alive, currently in the hospital, or currently residing in a skilled care facility) or because we could not identify a family caregiver. Fifty-three caregiver and older family member pairs provided informed consent. Participation, then, was 18.9% (53/(190+53+38)) if we include people not reached and 21.8% (53/(190+53) otherwise. Over 7 rounds of outreach (40–46 potential participants earmarked for contact each for a period of 2–3 weeks), the proportion providing consent was consistent. Fifty-three were randomized after completing baseline assessments, and 50 provided data at the follow-up assessment. The 3 not completing follow-up had transitioned to long-term care settings (1 intervention, 1 control) or were lost to follow-up (1 control). 25 pairs were randomized to immediate access (intervention) and 25 to delayed (control).
Psychometric Properties of Caregiver-Reported Measures
The 15 positive affect items formed a highly reliable single factor scale with an alpha coefficient of 0.92. The 5 negative affect items similarly formed a highly reliable scale with a coefficient of 0.87. Scale reliability for the Companion Animal Attachment Scale was 0.90.
Status of Participants at Baseline
Table 1 shows the status of family caregivers by treatment allocation as well as the dementia status of older adults participating in the trial. At baseline, the 2 groups were well matched except for age. Caregivers in the delayed-pet control condition were younger. They did not differ in other sociodemographic indicators, level of disability, common health symptoms, or mental health. They also did not differ in coresidence with older adult care receivers, intensity of caregiving activity, or perceived stress associated with providing care.
TABLE 1.
Baseline Status of Family Caregivers, By Treatment Allocation
| Delayed Robot Pet, n = 25 | Immediate Robot Pet, n =25 | Test Statistic | df | |
|---|---|---|---|---|
| Demography | ||||
| Age (mean, SD) | 56.3(11.5) | 62.2 (8.9) | 2.05a | 48 |
| Race, n (%) | 0.98 | 2 | ||
| White | 20 (83.0) | 20 (80.0) | ||
| Asian | 2 (8.3) | 1 (4.0) | ||
| African American | 2 (8.3) | 4 (16.0) | ||
| Female, n (%) | 19 (76.0) | 17 (68.0) | 0.40 | 1 |
| <College education, n (%) | 8 (32.0) | 5 (20.0) | 0.94 | 1 |
| <$50K annual income, n (%) | 9 (36.0) | 10 (40.0) | 0.08 | 1 |
| Pet in home, n (%) | 16 (64.0) | 17 (68.0) | 0.09 | 1 |
| Health | ||||
| Fair-poor health, n (%) | 6 (24.0) | 1 (4.0) | 4.15b | 1 |
| Much difficulty/unable to … | ||||
| Do housework, n (%) | 2 (8.0) | 0 (0.0) | 2.08 | 1 |
| Walk 15 min, n (%) | 3 (12.0) | 2 (8.0) | 0.22 | 1 |
| Do errands, n (%) | 2 (8.0) | 0 (0.0) | 2.08 | 1 |
| Past 7 days, often … | ||||
| Fatigued, n (%) | 6 (24.0) | 4 (16.0) | 0.50 | 1 |
| Sleep problems, n (%) | 11 (46.0) | 16 (64.0) | 2.01 | 1 |
| Pain, n (%) | 3 (12.0) | 1 (4.0) | 1.09 | 1 |
| “Brain fog,” n (%) | 1 (4.0) | 1 (4.0) | 0 | 1 |
| Overwhelming worry, n (%) | 4 (16.0) | 3 (12.0) | 0.17 | 1 |
| Caregiving | ||||
| Co-resident, n (%) | 10 (40.0) | 9 (36.0) | 0.85 | 1 |
| 2+ years caregiver, n (%) | 25 (100.0) | 21 (84.0) | 4.35b | 1 |
| >20 hr/wk care, n (%) | 14 (56.0) | 13 (52.0) | 0.08 | 1 |
| Other caregivers, n (%) | 9 (36.0) | 5 (20.0) | 1.59 | 1 |
| Provide ADL care, n (%) | 5 (20.0) | 9 (36.0) | 1.59 | 1 |
| Often frustrated by caregiving tasks, n (%) | 9 (36.0) | 11 (44.0) | 0.33 | 1 |
| Care receiver, Dementia status | ||||
| Dementia diagnosis, n (%) | 10 (40.0) | 14 (56.0) | 1.28 | 1 |
| Quick dementia rating, n (%) | 2.52 | 4 | ||
| Normal | 3 (12.0) | 1 (4.0) | ||
| Mild cognitive impairment | 6 (24.0) | 5 (20.0) | ||
| Mild dementia | 7 (20.0) | 11 (40.0) | ||
| Moderate dementia | 6 (24.0) | 4 (16.0) | ||
| Severe dementia | 3 (12.0) | 4 (16.0) | ||
| Reported affect, care receivers | ||||
| Positive (mean, SD) | 43.7 (10.6) | 40.7 (11.9) | 0.93 | 48 |
| Negative (mean, SD) | 12.8 (3.4) | 11.5 (3.7) | 1.36 | 48 |
Categorical measures assessed with X2; Fisher’s exact test used for cross-tabulations with cells <5. Continuous measures assessed by t-test.
p < 0.05 by t-test (age).
Fisher’s exact test: fair-poor health p = 0.10; 2+ years as caregiver, p = 0.11.
The proportion of older adults with dementia in the 2 groups was also similar. About half of the caregivers reported a diagnosis. Likewise, the distribution of dementia severity in the 2 study arms was similar, as shown in computed scores from the interviewer-administered Quick Dementia Rating Scale.
Delivery of the Intervention and Timing of Assessments
After consent and completion of the baseline survey, participants randomized to the immediate pet-access arm received the pet within 2–3 weeks. These participants had access to the pet for a mean (SD) of 32.9 (17.6) days before the follow-up interview. After completing the follow-up interview, participants in the control group received pets.
Outcomes
Engagement with Joy for All Pets
Caregivers in the immediate-access intervention arm completed the pet therapy engagement measure as part of the follow-up interview, which allowed them to report on the older person’s engagement with the robot pet. Caregivers reported a mean (SD) of 16.9 (6.1) in the 6-item involvement with pet scale (range 6–30).
Twenty-two of the 25 in the intervention arm provided open-ended comments about care receivers’ engagement with the pets. Nearly three-quarters, 16, reported positive engagement; 6 reported that older adults responded negatively or ignored the pet. Negative comments mostly involved lack of interest or inability to engage with the pet, that is, “his experience with the pet is basically no because he doesn’t realize it’s there, really.”
Effect on Positive and Negative Affect
At baseline, intervention and control arms did not significantly differ in the caregiver-reported positive and negative affect measures (Table 1). Mean (SD) positive affect was 40.7 (11.9) in the intervention group and 43.7 (10,6) in the control condition. Negative affect was 11.5 (3.7) in the intervention and 12.8 (3.8) in the control.
As shown in Figure 1, over the approximately 1 month of follow-up, positive affect scores increased from 40.7 to 45.9 in the intervention arm and declined from 43.7 to 41.6 in the control arm. The figure shows jittered points for each caregiver report as well as fitted regression lines. In a mixed model that included dementia severity score, involvement in ADL care, and caregiver age, positive affect, indicated by the group * time interaction term, increased by 7.28 points (, z = 3.36, p <0.001). The mixed model is shown in Table 2. Greater severity of dementia was an independent predictor of declines in positive affect (−5.1 points, z= −4.31, p <0.001).
FIGURE 1.
Change in positive affect by pet assignment.
TABLE 2.
Mixed Model Estimate of Effect of Treatment on Positive Affect
| B | se | z | 95% CI | |
|---|---|---|---|---|
| Caregiver age | −0.02 | 0.12 | −0.11 | −0.25, 0.26 |
| Dementia severity | −5.07 **** | 1.12 | −4.31 | −7.2, −2.9 |
| Provide regular ADL care | 0.90 | 2.76 | 0.31 | −4.5, 6.3 |
| Treatment group (baseline) | −2.00 | 2.89 | −0.66 | −7.7, 3.7 |
| Time | −2.12 | 1.50 | −1.39 | −5.1, 0.8 |
| Group * Time | 7.28*** | 2.12 | 3.36 | 3.1, 11.4 |
| Constant | 58.6 | 8.2 | 6.76 | 42.5, 74.8 |
Model Wald X2, =39.1, df = 6, p <0.0001 (Log likelihood −348.9).
p <0.001,.
p <0.0001.
As a simpler test, we examined the proportion of people in each group with high positive affect scores at follow-up. A score of 50 or above in positive affect represents the top quartile of caregiver reports. In the intervention arm, 44% (11/25) had scores of 50 or higher. In the control group, only 12% (3/25) met this threshold (p = 0.012, X2 = 6.35, 1 df). The full distribution of positive affect scores at follow-up is shown as overlapping histograms in Figure 2. At baseline, the 2 study arms did not differ in the proportion with scores of 50 or higher (control, 32%; intervention, 28%).
FIGURE 2.
Distribution of positive affect, follow up.
Intervention and control arms did not differ in changes in negative affect in mixed model analyses. Over the month of follow-up, negative affect scores declined from 11.4 to 10.8 in the intervention arm, and from 12.8 to 12.7 in the control arm.
Involvement with Robot Pet and Effects on Mood
Pet involvement was significantly correlated with positive affect, Pearson r = 0.43, p <0.035, df = 20 (1 participant missing data on pet involvement), as shown in Figure 3. Negative affect was not associated with pet involvement.
FIGURE 3.
Association of positive affect and pet involvement: intervention arm.
Caregivers described positive interaction with the pets. As one noted, “She would pet the dog and pull on its ears, pull on its tail. Snuggle it up to her and smile when it barked. Whenever she’s in the recliner or in the bed, I would give it to her and she would perk up a bit and hold it, or sometimes when it would make a noise she would smile or grin.” This caregiver went on to say, “I can hand the dog to her and get it started with making a noise, and she will smile and hug it a bit and start to treat it like a real dog. Sometimes it’s the only real response I get from her at this stage.”
Even when older adults did not engage with pets, simply having the pets promoted social engagement with others. As one caregiver pointed out, “She doesn’t engage with it, but she really loves showing it off when she gets visitors. She gets really excited about showing it to other people. She says, “talk to it, see how it moves!” She really enjoys seeing how others like how it moves around, barks, moves its tail. … She enjoys seeing other people enjoy the puppy.”
DISCUSSION
This pilot shows that a low-cost robot pet can improve positive affect in older people receiving family care. The rating of positive affect comes from caregivers, rather than reports from the older person receiving care or third-party observers. Because of the telephonic pilot nature of the research, we did not collect data in homes or interview older people themselves. Still, older people were consented as part of the research contact (a requirement from our IRB) and could decline participation.
While it is tempting to attribute the gain in positive affect to interaction with the pet, comments from caregivers suggest something else, or in addition, may be at work. The pet changes home environments. For example, caregivers report that the pets are something to show off and talk about when people visit. As one caregiver reported, “he is showing the neighbors and everyone that comes in his new cat …. This was a very interesting thing.” Caregivers in some cases reported that older people enjoyed watching how others interact with the pet even if they themselves did not engage with it.
Other studies have found similar environmental stimulation with exposure to the pets. For example, a study involving PARO noted that “the companion robot encouraged positive interactions with visitors and other family members. Caregivers highlighted how Paro served as a ‘good talking point’ and care recipients ‘enjoyed showing Paro off’.”14 The robot pets can be a topic of conversation and in this way increase social interaction, which in turn may improve mood. Thus, the pets should be considered an environmental stimulus as much as an interaction partner.
Limitations of the study include small sample size. However, the mixed model, which adjusted for caregiver age and involvement in caregiving as well as dementia severity, showed increases in positive affect. Older adults in the intervention arm were more likely to be in the top quartile of positive affect at follow-up. We were also limited in outcome measures. We adapted the PROMIS scales for proxy reports. While internal reliability for the positive and negative scales was high, proxy response departs from usual administration and should not be compared to PROMIS norms.
The research suggests a need for larger randomized trials with more careful consideration of control conditions. Rather than simply compare pet exposure to a null control condition, as implemented in this study, it may be valuable to compare the effect of the robot pet to a simple stuffed animal. This would isolate the effect of the robot pet’s greater potential as a partner for engagement. Or if environmental change is the key mechanism for positive affect and interaction, the robot pet could be compared to other changes in the home environment, such as introduction of plants or prescribed time spent with visitors.
To conclude, this randomized pilot trial suggests psychosocial benefit for the low-cost Joy for All robot pet companion.
Highlights.
-
What is the primary question addressed by this study?
Is exposure to a low-cost, robot companion pet associated with improvement in positive affect and reduction in negative affect in people receiving family caregiving support?.
-
What is the main finding of this study?
This randomized controlled trial suggests psychosocial benefit for the low-cost Joy for All robot pet companion. Older adults in the intervention arm were significantly more likely to be in the top quartile of positive affect at follow-up.
-
What is the meaning of the finding?
The mechanism for this benefit may involve a change in environment as well as direct interaction with the robot pet companion.
Footnotes
DISCLOSURE
The study was funded through internal funds from the Department of Behavioral and Community Health Sciences, School of Public Health, University of Pittsburgh. Dr. Gilliam is Director of the Area Agency on Aging, Allegheny County Department of Human Services. The agency provided the Joy for All Pets. For the remaining authors none were declared.
DATA STATEMENT
Data, analytic methods, and materials are available to other researchers for replication purposes and can be accessed by contacting the corresponding author. The study was pre-registered in clinicaltrials.gov. The data has not been previously presented orally or by poster at a scientific meeting.
References
- 1.Eachus P Pets, people and robots: the role of companion animals and robopets in the promotion of health and well-being. Int J Health Promotion Educ 39(1): 7–13. doi: 10.1080/14635240.2001.10806140 [DOI] [Google Scholar]
- 2.Bartneck C, Forlizzi J A design-centred framework for social human-robot interaction. 2004 presented at: RO-MAN 2004. 13th IEEE International Workshop on Robot and Human Interactive Communication (IEEE Catalog No.04TH8759). 10.1109/roman.2004.1374827 [DOI] [Google Scholar]
- 3.Hudson J, Ungar R, Albright L, et al. : Robotic pet use among community-dwelling older adults. J Gerontol B Psychol Sci 2020; 75 (9):2018–2028;doi: 10.1093/geronb/gbaa119 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Koh WQ, Ang FXH, Casey D: Impacts of low-cost robotic pets for older adults and people with dementia: scoping review. JMIR Rehabil Assist Technol 2021; 8(1):e25340;doi: 10.2196/25340 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bradwell HL, Edwards KJ, Winnington R, et al. : Companion robots for older people: importance of user-centred design demonstrated through observations and focus groups comparing preferences of older people and roboticists in South West England. BMJ Open 2019; 9:e032468;doi: 10.1136/bmjopen-2019-032468 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Pike J, Picking R, Cunningham S: Robot companion cats for people at home with dementia: a qualitative case study on companotics. Dementia (London) 2020; 16:1471301220932780; doi: 10.1177/1471301220932780 [DOI] [PubMed] [Google Scholar]
- 7.Harris-Gersten ML, Davagnino JM, Alcorn ER, et al. : Usability and acceptability of social robot pets among community-dwelling veterans living with dementia and their caregivers. Am J Alzheimers Dis Other Demen 2023; 38:1–7;doi: 10.1177/15333175231200973 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wexler S, Drury L, Pollak C, et al. : Robotic pet therapy improves outcomes for hospitalized older adults. Innov Aging 2018; 2 (1):731 [Google Scholar]
- 9.Lu LC, Lan SH, Hsieh YP, et al. : Effectiveness of companion robot care for dementia: a systematic review and meta-analysis. Innov Aging 2021; 5(2):igab013;doi: 10.1093/geroni/igab013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Moyle W, Cooke M, Beattie E, et al. : Exploring the effect of companion robots on emotional expression in older adults with dementia: a pilot randomized controlled trial. J Gerontol Nursing 2013; 39(5):46–53;doi: 10.3928/00989134-20130313-03 [DOI] [PubMed] [Google Scholar]
- 11.Petersen S, Houston S, Qin H, et al. : The utilization of robotic pets in dementia care. J Alzheimers Dis 2017; 55(2):569–574; doi: 10.3233/JAD-160703 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Mervin MC, Moyle W, Jones C, et al. : The cost-effectiveness of using PARO, a therapeutic robotic seal, to reduce agitation and medication use in dementia: findings from a cluster–Randomized controlled trial. J Am Med Directors Assoc 2018; 19(7), 619–622.e1 [DOI] [PubMed] [Google Scholar]
- 13.Kang SY, Kim J, Kuhl M: The impact of companion robotic pets on well-being among community-dwelling women aged 65 and older. Innov Aging 2024; 8(Supplement 1):900;doi: 10.1093/ger-oni/igae098.2908 [DOI] [Google Scholar]
- 14.Liang A, Piroth I, Robinson H, et al. : A pilot randomized trial of a companion robot for people with dementia living in the community. J Am Med Dir Assoc 2017; 18(10):871–878;doi: 10.1016/j.jamda.2017.05.019 [DOI] [PubMed] [Google Scholar]
- 15.Galvin JE: The Quick Dementia Rating System (QDRS): a rapid dementia staging tool. Alzheimers Dement (Amst) 2015; 1 (2):249–259;doi: 10.1016/j.dadm.2015.03.00316 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Poresky RH, Hendrix C, Mosier JE, et al. : The companion animal bonding scale: internal reliability and construct validity. Psychol Rep 1987; 60(3):743–746;doi: 10.2466/pr0.1987.60.3.743 [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data, analytic methods, and materials are available to other researchers for replication purposes and can be accessed by contacting the corresponding author. The study was pre-registered in clinicaltrials.gov. The data has not been previously presented orally or by poster at a scientific meeting.