Abstract
This feasibility study explored the impact of fostering a shelter cat on loneliness and well-being in older adults living alone without a pet. The study also examined the effect of cat fostering on older adults’ interest in cat adoption when perceived barriers to adoption were removed.
A total of 29 adults (age ≥60) were paired with a cat and asked to foster for a minimum of 4 months (with an option to adopt). Participants completed surveys before placement with cats, with a follow-up at 1 and 4 months postplacement and 12 months if they adopted their cat. The survey included scales of loneliness, physical and mental health, self-efficacy, positive and negative affect, and comfort from a companion animal. Scores were compared across time using a Multivariable Linear Mixed Model.
A total of 23 of 29 (79.3%) participants remained in the study for at least 4 months, and a majority (95.7%) adopted their cat. Differences in marginal means (adjusted for physical health) showed a significant improvement in loneliness from baseline to 4 months (p = .029). A similar 4-month improvement that approached statistical significance (adjusted p = .079) was observed for mental health. No other scales showed statistically significant changes across time.
Fostering a shelter cat with the option for adoption may be an effective solution for alleviating loneliness and improving mental health in older adults. Interest in adopting foster cats was high when perceived barriers to adoption were removed.
The proportion of older adults living alone is rising in the United States. In 2021, approximately 27% of adults age 65 or older lived alone, and 43% of women age 75 or older lived alone (The Administration for Community Living, 2022). Living alone can be a risk factor for multiple health outcomes, including increased mortality risk and loneliness. Although living alone does not necessarily cause loneliness in older adults, it is a major risk factor for loneliness. Loneliness and living alone increase the likelihood of premature mortality by 26% and 32%, respectively (Holt-Lunstad et al., 2015). Loneliness is also associated with several physical and mental health diseases (Holt-Lunstad et al., 2015; Rico-Uribe et al., 2018). In 2017, the American Association of Retired People estimated the cost to Medicare in the United States associated with loneliness in older adults to be $6.7 billion annually (AARP, 2017). More research is needed to address the causes and amelioration of loneliness in older adults.
Studies evaluating companion animals’ effects on loneliness in older adults show mixed results, with some demonstrating decreases in loneliness while others detected no significant impact on loneliness. Systematic reviews of this literature identify potential limitations associated with many of these studies, including most being cross-sectional and challenges related to conducting randomized controlled trials in this area of research (Gilby & Tani, 2015; Kretzler et al., 2022). Few studies have evaluated pets’ long-term effects on older adults living independently in a home environment, and even fewer studies have involved cats. The authors of these reviews recommend additional longitudinal studies evaluating the impact of pets on loneliness and social isolation; and that such studies should examine the effects of new pet ownership in older adults without a pet at baseline.
The majority of studies showing the benefits of pet ownership in older adults are cross-sectional studies, and most fall into one of the following categories: community-dwelling older adults who already have a pet (Hui Gan et al., 2020); animal-assisted interventions for a short period of time in nursing homes or long-term care facilities, for example, a pet is brought in to a nursing home and residents interact with it for a few hours (Banks & Banks, 2002); or comparisons of older adults who already have a pet with non-pet owners (Branson et al., 2016; Friedmann et al., 2020; Heuberger, 2017). Research is needed to assess the long-term impact of companionship from a pet on community-dwelling older adults living alone before and after receiving a pet. Further understanding of the role pets have as an intervention for loneliness and physical and mental health in this population is needed to better understand the role pets have in helping older adults age in place and to support policies that facilitate pet ownership.
The recent coronavirus disease 2019 (COVID-19) pandemic further illustrates the impact of loneliness and social isolation have on older adults’ well-being. A limited number of peripandemic studies evaluated the effects pets had on loneliness in adults, with mixed results (Kretzler et al., 2022). One study examining the impact of dog or cat companionship on loneliness found no benefits (Olivia & Johnston, 2020), while another study showed decreased loneliness with the presence of dogs and cats combined (Kogan et al., 2021). A third study examining the effect of cat companions showed the owners were less likely to feel socially isolated but not less lonely (Kogan et al., 2021). None of these studies focused solely on older adults, and all were cross-sectional in design, further illustrating the need for more longitudinal studies in this area of research.
The health benefits of companion animals have been widely reported (Friedman & Krause-Parello, 2018). Several prior studies have shown the beneficial effects of companion dogs and cats on physical and mental health in older adults specifically (Enmarker et al., 2012, 2015). These have included decreased social isolation and depressive symptoms, increased well-being, reduced mortality, decreased cardiovascular risk factors (dogs), and increased physical exercise (Friedmann et al., 1980).
Although a majority of studies show positive effects of animal companionship, the results are conflicting when the focus is on cats specifically. Although there are far fewer studies focusing specifically on cats, many show no significant difference in loneliness for cat companionship compared to dogs or no pets (Kretzler et al., 2022).
Although the need exists to better address loneliness in older adults, at the same time, approximately 6.3 million homeless dogs and cats end up in shelters every year in the United States (ASPCA, 2019). As some shelters strive to become low-kill/no-kill shelters, foster homes are needed to bridge the time from pet intake to adoption into a permanent home. If older adults living alone foster shelter cats, this could benefit both the older adults and homeless cats.
Despite the potential for mutual benefit, older adults may face barriers to adopting a cat or may be reluctant to adopt one for various reasons (Anderson et al., 2015). First, some older adults have never lived with an indoor cat and are uncertain whether a cat is the correct pet for them. Second, older adults may be concerned about what happens to their cats if they become ill or pass away. Third, some older adults struggle with the financial responsibilities associated with pet ownership. Finally, some older adults have health issues, such as limited mobility, that affect their ability to care for a pet. Given these factors, a successful cat-fostering program requires leveraging strong partnerships between older adults and community organizations.
This pilot study addresses previous investigators’ call for longitudinal research while leveraging academic and community partnerships. The objectives of this feasibility study were to (1) determine if fostering a shelter cat, with the option for adoption, improves loneliness, mental and physical health of older adults living alone and without other pets, and (2) determine the impact fostering a cat had on older adults’ interest in and commitment to adopting their foster cat when barriers were removed.
Method
Institutional Approval
The part of the study working with older adults was approved by The University of Georgia Institutional Review Board (#STUDY00005341). The use of cats in the study was approved by The University of Georgia College of Veterinary Medicine Clinical Research Committee. Both animal organizations supplying cats for the study signed consent forms allowing their cats to participate in the study.
Participants
Eligible participants were recruited through flyers posted at The Athens Community Council of Aging, regional community organizations, businesses, local media stories about the study, and other senior community centers within a 70-mile radius of Athens, GA, USA. Recruitment of human participants was challenging, partly because the target population was older adults living alone. The challenges increased dramatically once the pandemic reached the United States. However, the study enrolled 29 older adults who match the following inclusion criteria: (1) age ≥ 60 years, (2) living alone, (3) not currently a pet owner, (4) willing to foster an indoor-only cat, (5) residing in a community residence allowing pets, (6) access to a telephone, and (7) having an acceptable and safe environment for a cat to live (e.g., absence of poisonous plants, etc.). Age 60 was chosen as the cut-off to match many national pet fostering programs whose eligibility is age 60 (e.g., Pets for the Elderly Foundation, 2023), as well as to broaden our reach to a wider audience In addition, participants were assessed for their functional and cognitive ability, with exclusion criteria including ≥2 deficits in Activities of Daily Living (Wade & Collin, 1988), scoring below the cut-off (answering more than three of 23 questions incorrectly) on the Adult Lifestyles and Function Interview—Mini-Mental State Exam (Fischbach, 1990; Roccaforte et al., 1992), or having fallen during the past year. A fall is defined as an event that results in coming to rest inadvertently on the ground or lower level (World Health Organization, 2021).
Protocol
Enrollment
Potential participants were administered a telephone screening and, if eligible, consented to participate over the phone. The screening included outlining the risks of COVID-19 for those participants who entered the study after the pandemic began. All eligible participants who consented to the study were scheduled for an in-home visit with an investigator (strict COVID protocols were followed). Each eligible participant completed a baseline questionnaire administered over the phone or during the first home visit.
During initial home visits, participants signed forms for study consent and fostering agreement with the shelter. The home environment was assessed to ensure it was suitable and safe for a cat. Arrangements were made for payment of any required pet deposits. Participants were given copies of all forms, descriptions of the study team, information on cat-proofing their home, caring for their foster cat, handling and petting their cat, tips for preventing falls in their home, wound management for cat scratches or bites, and who to contact for urgent issues. Educational information about COVID-19 was added to the folder post-COVID.
Foster cat pairing
Participants chose their cat to foster to enhance the connection between the participant and their cat. Before the pandemic, an investigator brought participants to the shelter to select their cats. Postpandemic, the investigator went to the shelter and sent pictures and videos to the participant or called to describe the available cats. All cats came spayed or neutered, vaccinated and screened for Feline leukemia virus/Feline immunodeficiency virus (FeLV/FIV), dewormed, prophylactically treated for fleas, and microchipped. The microchip number for the cat also served as the participants’ identification number for the study.
Once participants were paired with their foster cats, the study provided all start-up supplies needed, including litter boxes, cat litter, scratching posts, carriers, bowls, cat toys, and cat food. Participants were assigned to an investigator who was a member of the Participant Support Team. Team members called participants monthly to check on recent falls and screen for any physical or cognitive decline. In addition, the study veterinarian conducted monthly home visits to deliver cat litter and food, trim nails, and ensure the cat was receiving good care. If veterinary medical care were needed, this same individual would address the problem or take the cat to a partner veterinary clinic for tests at no cost to participants.
Study protocol
Participants were asked to foster their cat for a minimum of 4 months; however, they could return their foster cat at any time. Participants could adopt their foster cat between 1 and 4 months during the fostering period. If they decided not to adopt by 4 months, the fostering period and study ended, and the cat was returned. If they chose to adopt their cat, the study paid the adoption fee, and the participants continued the study for up to 12 months. During this time, monthly phone calls from a member of the participant support team continued, as well as monthly home visits from the study veterinarian to assess the cat’s well-being and supply cat food and litter. Participants were told that if their living circumstances changed and they could no longer keep their cat or their cat outlived them, the cat could be returned to the shelter.
Participants completed a survey at baseline before being paired with their foster cat. The same follow-up survey, but with the addition of a scale to assess their comfort level with their foster cat, was administered 1 and 4 months into the fostering period. The survey was administered by phone and then in person during monthly home visits. If participants chose to adopt their foster cat, an additional survey was administered 12 months after they were paired with their cat.
The final analytic sample consisted of 23 participants aged 60–90 years (mean age 71.0, standard deviation 7.5) who completed the minimum enrollment time (4 months). The majority (83%) of the participants were female. The adopted cats ranged from 2 to 132 months old (mean 18.6 months, standard deviation 28.6) at the time of placement.
Measurements
Each person was asked the same questions in 1-, 4-, and 12-month follow-up surveys. Loneliness was assessed using the Revised UCLA three-item loneliness scale (Hughes et al., 2004). Respondents rate on a 3-point scale (1 = hardly ever, 2 = some of the time, 3 = often) how often they lack companionship, feel left out, and feel isolated. Scores are then summed, with greater scores indicating greater levels of loneliness. Mental and physical health was assessed using the Short Form-12 (SF-12) Health Survey version 1.0, which provides a composite score for both mental and physical health (Ware et al., 1996). A mental component score and a physical component score were calculated using an online orthotoolkit (John Ware Research Group, Inc., 2009); higher scores indicate better health. Perceived self-efficacy was assessed using The Generalized Self-Efficacy Scale (Schwarzer & Jerusalem, 1995). This is a 10-item scale, and respondents rate on a 4-point scale (1 = not true at all to 4 = exactly true) how well they agree with 10 statements about their coping ability after experiencing various stressful life events. Composite scores range from 10 to 40, with higher scores indicating greater perceived self-efficacy. Effect was measured using the Positive and Negative Affect Scale (Watson et al., 1988), modified to a 12-item scale following previous research (Mroczek & Kolarz, 1998). Each item asks, “during the past 30 days, how much of the time did you feel …” followed by six positive affect and six negative affect words or phrases. Respondents rate on a 5-point scale (1 = none of the time to 5 = all of the time). The scale is coded into positive and negative affect with a range of 10–50 for both items. For the positive affect score, higher scores indicate more positive affect, whereas for the negative affect score, lower scores indicate less negative affect.
Finally, after the baseline survey, each subsequent survey asked respondents about their perceived comfort resulting from their cat, using a modified 11-item Comfort from Companion Animal Scale (Zasloff, 1996). The word “cat” was replaced with “foster cat,” two items on safety and exercise that pertained to dogs were removed from the original 13-item scale, and respondents rated their level of agreement with a statement about their foster cat on a 5-point scale (1 = strongly disagree to 5 = strongly agree). A cumulative score was calculated, with higher scores indicating a greater perceived comfort level.
Statistical Analyses
Scores from the scales were compared between time points with a linear mixed model (LMM) to account for within-subject correlation. Histograms and Q–Q plots were examined and confirmed the assumption of normality except for comfort from the companion animal scale, which was analyzed with a Friedman test. LMMs with a single fixed factor of time and a random factor of subject were used to estimate and test differences between each time to baseline. Multivariable LMMs containing an additional covariate of the physical composite scale were developed to adjust for differences in physical health between and within subjects. Satterthwaite degrees of freedom method and restricted maximum likelihood (REML) estimation were used. As analyses were considered exploratory, no adjustments for multiple tests were made. All analyses were performed using SAS 9.4 (Cary, NC, USA) and a significance threshold of 0.05.
Data were collected from all outcomes of pairings of participants with their foster cats. Possible outcomes included whether participants chose to adopt their foster cat within 4 months, chose to return their foster cat, or were removed from the study. If participants chose to adopt their foster cat, the month during the fostering period of adoption was recorded. If participants chose to return their foster cat or were removed from the study, the reason was noted.
Results
Participants
Twenty-nine participants were paired with a cat during the study. Twenty-three of 29 participants (79.3%) completed the minimum time of enrollment (4 months) required for their data to be included in the statistical analysis. The majority (87%) of participants remained in the study for 12 months, and the remaining 13% completed at least 4 months. As this study began before the pandemic but continued enrollment throughout the pandemic, some survey results may be affected. Table 1 shows the enrollment of participants at different data collection points pre- and postpandemic.
Number (%) of Surveys Administered Pre- and Postpandemic at Different Data Collection Points in Older Adults Paired with a Cat
| Baseline | 1 month | 4 month | 12 month | |
|---|---|---|---|---|
| Pre-COVID surveys | 18 (78%) | 12 (57%) | 10 (45%) | 6 (30%) |
| Post-COVID surveys | 5 (22%) | 9 (43%) | 12 (55%) | 14 (70%) |
| Total no. of participants | 23 | 21 | 22 | 20 |
| Baseline | 1 month | 4 month | 12 month | |
|---|---|---|---|---|
| Pre-COVID surveys | 18 (78%) | 12 (57%) | 10 (45%) | 6 (30%) |
| Post-COVID surveys | 5 (22%) | 9 (43%) | 12 (55%) | 14 (70%) |
| Total no. of participants | 23 | 21 | 22 | 20 |
Number (%) of Surveys Administered Pre- and Postpandemic at Different Data Collection Points in Older Adults Paired with a Cat
| Baseline | 1 month | 4 month | 12 month | |
|---|---|---|---|---|
| Pre-COVID surveys | 18 (78%) | 12 (57%) | 10 (45%) | 6 (30%) |
| Post-COVID surveys | 5 (22%) | 9 (43%) | 12 (55%) | 14 (70%) |
| Total no. of participants | 23 | 21 | 22 | 20 |
| Baseline | 1 month | 4 month | 12 month | |
|---|---|---|---|---|
| Pre-COVID surveys | 18 (78%) | 12 (57%) | 10 (45%) | 6 (30%) |
| Post-COVID surveys | 5 (22%) | 9 (43%) | 12 (55%) | 14 (70%) |
| Total no. of participants | 23 | 21 | 22 | 20 |
To determine if fostering a shelter cat improved outcomes, we ran LMMs to examine the main effect of time for each scale outcome, with a second model adjusting for physical scores as a covariate. The main effect of time was not statstically significant for any scale (based on unadjusted or adjusted p values). However, when the loneliness scale was adjusted for physical score, it approached statistical significance (p = .084; data not shown).
Table 2 shows the differences in marginal means (adjusted for missing values when subjects failed to complete the survey each time). There are two models, (1) the unadjusted models containing times as the fixed factor and (2) the adjusted model that includes physical health composite score as a covariate. There was a significant change (decrease) in loneliness score of −1.4 (95% CI: −2.6 to −0.1) between baseline and 4 months when adjusted for physical health (though not based on unadjusted p value = .163). At month 12 the loneliness score p value increased (.354), no longer significant.
Unadjusted and Adjusted Mean Differences from Baseline after Pairing Older Adults with a Cat
| Scale | Time (months) | Unadjusted mean differencea (95% CI) from baseline | Unadjusted p value | Adjusted mean differenceb (95% CI) from baseline | Adjusted p value |
|---|---|---|---|---|---|
| UCLA loneliness scale | 1 | −0.4 (−1.7 to 0.9) | .819 | −0.6 (−1.9 to 0.6) | .496 |
| UCLA loneliness scale | 4 | −1.0 (−2.4 to 0.3) | .163 | −1.4 (−2.6 to −0.1) | .029c |
| UCLA loneliness scale | 12 | −0.5 (−1.8 to 0.9) | .770 | −0.8 (−2.0 to 0.5) | .354 |
| Generalized self−efficacy scale | 1 | 0.1 (−3.4 to 3.7) | 1.000 | 0.7 (−2.5 to 4.0) | .903 |
| Generalized self−efficacy scale | 4 | −0.1 (−3.6 to 3.4) | 1.000 | 0.8 (−2.4 to 4.1) | .875 |
| Generalized self-efficacy scale | 12 | 0.3 (−3.3 to 3.9) | .995 | 1.2 (−2.2 to 4.5) | .736 |
| Physical health | 1 | −3.5 (−10.7 to 3.8) | .534 | ||
| Physical health | 4 | −5.1 (−12.3 to 2.1) | .227 | ||
| Physical health | 12 | −4.9 (−12.2 to 2.5) | .273 | ||
| Mental health | 1 | 2.1 (−4.3 to 8.5) | .768 | 2.6 (−3.7 to 9.0) | .629 |
| Mental health | 4 | 5.1 (−1.3 to 11.4) | .146 | 5.8 (−0.5 to 12.2) | .079 |
| Mental health | 12 | 3.0 (−3.5 to 9.5) | .549 | 3.7 (−2.8 to 10.2) | .377 |
| Positive affect | 1 | 0.4 (−2.2 to 3.1) | .962 | 0.8 (−1.7 to 3.4) | .769 |
| Positive affect | 4 | −1.2 (−3.9 to 1.4) | .551 | −0.6 (−3.2 to 1.9) | .884 |
| Positive affect | 12 | 0.0 (−2.7 to 2.7) | 1.000 | 0.6 (−2.0 to 3.2) | .912 |
| Negative affect | 1 | −0.7 (−3.1 to 1.8) | .858 | −1.1 (−3.4 to 1.2) | .512 |
| Negative affect | 4 | −0.6 (−3 to 1.9) | .892 | −1.3 (−3.6 to 1.0) | .412 |
| Negative affect | 12 | 0.4 (−2.1 to 2.9) | .966 | −0.2 (−2.6 to 2.1) | .990 |
| Scale | Time (months) | Unadjusted mean differencea (95% CI) from baseline | Unadjusted p value | Adjusted mean differenceb (95% CI) from baseline | Adjusted p value |
|---|---|---|---|---|---|
| UCLA loneliness scale | 1 | −0.4 (−1.7 to 0.9) | .819 | −0.6 (−1.9 to 0.6) | .496 |
| UCLA loneliness scale | 4 | −1.0 (−2.4 to 0.3) | .163 | −1.4 (−2.6 to −0.1) | .029c |
| UCLA loneliness scale | 12 | −0.5 (−1.8 to 0.9) | .770 | −0.8 (−2.0 to 0.5) | .354 |
| Generalized self−efficacy scale | 1 | 0.1 (−3.4 to 3.7) | 1.000 | 0.7 (−2.5 to 4.0) | .903 |
| Generalized self−efficacy scale | 4 | −0.1 (−3.6 to 3.4) | 1.000 | 0.8 (−2.4 to 4.1) | .875 |
| Generalized self-efficacy scale | 12 | 0.3 (−3.3 to 3.9) | .995 | 1.2 (−2.2 to 4.5) | .736 |
| Physical health | 1 | −3.5 (−10.7 to 3.8) | .534 | ||
| Physical health | 4 | −5.1 (−12.3 to 2.1) | .227 | ||
| Physical health | 12 | −4.9 (−12.2 to 2.5) | .273 | ||
| Mental health | 1 | 2.1 (−4.3 to 8.5) | .768 | 2.6 (−3.7 to 9.0) | .629 |
| Mental health | 4 | 5.1 (−1.3 to 11.4) | .146 | 5.8 (−0.5 to 12.2) | .079 |
| Mental health | 12 | 3.0 (−3.5 to 9.5) | .549 | 3.7 (−2.8 to 10.2) | .377 |
| Positive affect | 1 | 0.4 (−2.2 to 3.1) | .962 | 0.8 (−1.7 to 3.4) | .769 |
| Positive affect | 4 | −1.2 (−3.9 to 1.4) | .551 | −0.6 (−3.2 to 1.9) | .884 |
| Positive affect | 12 | 0.0 (−2.7 to 2.7) | 1.000 | 0.6 (−2.0 to 3.2) | .912 |
| Negative affect | 1 | −0.7 (−3.1 to 1.8) | .858 | −1.1 (−3.4 to 1.2) | .512 |
| Negative affect | 4 | −0.6 (−3 to 1.9) | .892 | −1.3 (−3.6 to 1.0) | .412 |
| Negative affect | 12 | 0.4 (−2.1 to 2.9) | .966 | −0.2 (−2.6 to 2.1) | .990 |
Notes: CI = confidence interval
aDifference in estimated marginal means.
bAdjusted for physical composite score.
cStatistically significant result
Unadjusted and Adjusted Mean Differences from Baseline after Pairing Older Adults with a Cat
| Scale | Time (months) | Unadjusted mean differencea (95% CI) from baseline | Unadjusted p value | Adjusted mean differenceb (95% CI) from baseline | Adjusted p value |
|---|---|---|---|---|---|
| UCLA loneliness scale | 1 | −0.4 (−1.7 to 0.9) | .819 | −0.6 (−1.9 to 0.6) | .496 |
| UCLA loneliness scale | 4 | −1.0 (−2.4 to 0.3) | .163 | −1.4 (−2.6 to −0.1) | .029c |
| UCLA loneliness scale | 12 | −0.5 (−1.8 to 0.9) | .770 | −0.8 (−2.0 to 0.5) | .354 |
| Generalized self−efficacy scale | 1 | 0.1 (−3.4 to 3.7) | 1.000 | 0.7 (−2.5 to 4.0) | .903 |
| Generalized self−efficacy scale | 4 | −0.1 (−3.6 to 3.4) | 1.000 | 0.8 (−2.4 to 4.1) | .875 |
| Generalized self-efficacy scale | 12 | 0.3 (−3.3 to 3.9) | .995 | 1.2 (−2.2 to 4.5) | .736 |
| Physical health | 1 | −3.5 (−10.7 to 3.8) | .534 | ||
| Physical health | 4 | −5.1 (−12.3 to 2.1) | .227 | ||
| Physical health | 12 | −4.9 (−12.2 to 2.5) | .273 | ||
| Mental health | 1 | 2.1 (−4.3 to 8.5) | .768 | 2.6 (−3.7 to 9.0) | .629 |
| Mental health | 4 | 5.1 (−1.3 to 11.4) | .146 | 5.8 (−0.5 to 12.2) | .079 |
| Mental health | 12 | 3.0 (−3.5 to 9.5) | .549 | 3.7 (−2.8 to 10.2) | .377 |
| Positive affect | 1 | 0.4 (−2.2 to 3.1) | .962 | 0.8 (−1.7 to 3.4) | .769 |
| Positive affect | 4 | −1.2 (−3.9 to 1.4) | .551 | −0.6 (−3.2 to 1.9) | .884 |
| Positive affect | 12 | 0.0 (−2.7 to 2.7) | 1.000 | 0.6 (−2.0 to 3.2) | .912 |
| Negative affect | 1 | −0.7 (−3.1 to 1.8) | .858 | −1.1 (−3.4 to 1.2) | .512 |
| Negative affect | 4 | −0.6 (−3 to 1.9) | .892 | −1.3 (−3.6 to 1.0) | .412 |
| Negative affect | 12 | 0.4 (−2.1 to 2.9) | .966 | −0.2 (−2.6 to 2.1) | .990 |
| Scale | Time (months) | Unadjusted mean differencea (95% CI) from baseline | Unadjusted p value | Adjusted mean differenceb (95% CI) from baseline | Adjusted p value |
|---|---|---|---|---|---|
| UCLA loneliness scale | 1 | −0.4 (−1.7 to 0.9) | .819 | −0.6 (−1.9 to 0.6) | .496 |
| UCLA loneliness scale | 4 | −1.0 (−2.4 to 0.3) | .163 | −1.4 (−2.6 to −0.1) | .029c |
| UCLA loneliness scale | 12 | −0.5 (−1.8 to 0.9) | .770 | −0.8 (−2.0 to 0.5) | .354 |
| Generalized self−efficacy scale | 1 | 0.1 (−3.4 to 3.7) | 1.000 | 0.7 (−2.5 to 4.0) | .903 |
| Generalized self−efficacy scale | 4 | −0.1 (−3.6 to 3.4) | 1.000 | 0.8 (−2.4 to 4.1) | .875 |
| Generalized self-efficacy scale | 12 | 0.3 (−3.3 to 3.9) | .995 | 1.2 (−2.2 to 4.5) | .736 |
| Physical health | 1 | −3.5 (−10.7 to 3.8) | .534 | ||
| Physical health | 4 | −5.1 (−12.3 to 2.1) | .227 | ||
| Physical health | 12 | −4.9 (−12.2 to 2.5) | .273 | ||
| Mental health | 1 | 2.1 (−4.3 to 8.5) | .768 | 2.6 (−3.7 to 9.0) | .629 |
| Mental health | 4 | 5.1 (−1.3 to 11.4) | .146 | 5.8 (−0.5 to 12.2) | .079 |
| Mental health | 12 | 3.0 (−3.5 to 9.5) | .549 | 3.7 (−2.8 to 10.2) | .377 |
| Positive affect | 1 | 0.4 (−2.2 to 3.1) | .962 | 0.8 (−1.7 to 3.4) | .769 |
| Positive affect | 4 | −1.2 (−3.9 to 1.4) | .551 | −0.6 (−3.2 to 1.9) | .884 |
| Positive affect | 12 | 0.0 (−2.7 to 2.7) | 1.000 | 0.6 (−2.0 to 3.2) | .912 |
| Negative affect | 1 | −0.7 (−3.1 to 1.8) | .858 | −1.1 (−3.4 to 1.2) | .512 |
| Negative affect | 4 | −0.6 (−3 to 1.9) | .892 | −1.3 (−3.6 to 1.0) | .412 |
| Negative affect | 12 | 0.4 (−2.1 to 2.9) | .966 | −0.2 (−2.6 to 2.1) | .990 |
Notes: CI = confidence interval
aDifference in estimated marginal means.
bAdjusted for physical composite score.
cStatistically significant result
A similar 4-month improvement that approached statistical significance (adjusted p = .079) was observed for mental health. Participants showed an increase from 49 at baseline to 54.1 at month 4. The score at 12 months dropped to 52.1. This score was numerically higher than at baseline and no longer close to significant (adjusted p value .377). The remaining outcomes did not change significantly over time.
Correlations of age with each scale averaged over all time points were tested with Pearson’s correlation analyses (data not shown). Age did not have any impact on the parameters assessed.
To explore changes across time, Figures 1 and 2 show box and whisker plots for all outcomes where the mean changed across time (data not shown). Each box is drawn from the 25th percentile to the 75th percentile. The vertical line inside the box shows the location of the median, and the symbol shows the location of the mean.
(A) shows the effects of companionship from a cat on loneliness in older adults. The higher the score on the scale, the greater the degree of loneliness. (B) shows the effect of companionship from a cat on generalized-self efficacy in older adults. The higher the score, the greater the degree of self-efficacy. (C) shows the effects of companionship from a cat on Physical & Mental Health scores in older adults. Higher scores indicate better physical and mental health.
(A) shows the effect of cat companionship on Positive and Negative Affect. For the Negative Affect Scale in older adults, lower scores represent lower levels of negative affect. For the Positive Affect Scale, higher scores represent higher levels of positive affect. (B) shows the effect of companionship from a cat on the older adults’ comfort after being paired with a cat. The maximum score for this scale is 55. The higher the number, the greater participants perceived comfort after receiving their cat.
Results for changes across time showed no significance at any time point for the five assessments analyzed. However, during the first 4 months after participants were paired with a cat, the data for the UCLA Loneliness Scale and the SF12 Mental Health Composite Scale showed numeric improvement. However, this trend did not persist for 12 months.
The second study objective was to determine the impact fostering a cat has on older adults’ interest in and commitment to adopting their foster cat when barriers to adopting a cat were removed. Table 3 shows the outcome of the initial 29 cat and participant pairings. Participants were eligible to adopt their foster cat between 1 and 4 months after being paired with their cat. Most participants (79.3%) paired with a cat chose to remain in the study for at least 4 months. Four of 29 (13.8%) voluntarily returned their cat before 4 months, and two others were removed from the study. One was removed before adopting their cat because the cat bit them while trying to trim its nails, and the other person was removed from the study at 8 months because they tripped over the cat and fell. They were not injured. The majority (95.7%) of participants who remained in the study for 4 months decided to adopt their foster cat. Half of the participants chose to adopt their cat the first month (one) they were eligible to adopt. Most of the remaining participants decided to wait until the last month (four) they were eligible to adopt.
Total Number of Older Adults Paired with a Cat and Outcomes of Pairings
| Number of participants | Cat adoption status | Adoption occurrence | Number of participants | ||
|---|---|---|---|---|---|
| Paired with a cat | Returning cat before 4 months | Removed from study before 4 months | Participants remaining in study ≥ 4 months, N = 23 | Length of time before adoption, N = 22 | Adopting cat and returning cat by 12 months |
| 29 | 4a | 2b | 22 (adopting) 1 (not adopting)c | 1 month = 11 2 months = 2 3 months = 0 4 months = 9 | 1d |
| Number of participants | Cat adoption status | Adoption occurrence | Number of participants | ||
|---|---|---|---|---|---|
| Paired with a cat | Returning cat before 4 months | Removed from study before 4 months | Participants remaining in study ≥ 4 months, N = 23 | Length of time before adoption, N = 22 | Adopting cat and returning cat by 12 months |
| 29 | 4a | 2b | 22 (adopting) 1 (not adopting)c | 1 month = 11 2 months = 2 3 months = 0 4 months = 9 | 1d |
Note: aTime and reason for returning the cat: (1) 2 days—did not like the smell of cat litter and did not want an indoor cat; (2) 2 months—kitten too rambunctious; (3) 2 months—participant afraid of tripping over cat; and (4) 3 months—did not like having to keep door to sewing room closed to prevent cat from getting in room.
bReason for removing two participants from study before 4 months: (1) participant fell and broke their collar bone (fall was not related to cat); (2) cat bit participant on finger while trying to trim the cat’s nails. A third participant was removed from the study at 8 months after being paired with their cat because they tripped over the cat and fell. This participant had already adopted their cat and chose to keep the cat.
cOne participant that fostered their cat for 4 months and chose not to adopt it went into the study genuinely interested in fostering a cat and helping to make it more adoptable.
dParticipant vacillated over whether to keep their cat or not. After the lockdown associated with COVID-19 ended, participant decided to return their cat.
Total Number of Older Adults Paired with a Cat and Outcomes of Pairings
| Number of participants | Cat adoption status | Adoption occurrence | Number of participants | ||
|---|---|---|---|---|---|
| Paired with a cat | Returning cat before 4 months | Removed from study before 4 months | Participants remaining in study ≥ 4 months, N = 23 | Length of time before adoption, N = 22 | Adopting cat and returning cat by 12 months |
| 29 | 4a | 2b | 22 (adopting) 1 (not adopting)c | 1 month = 11 2 months = 2 3 months = 0 4 months = 9 | 1d |
| Number of participants | Cat adoption status | Adoption occurrence | Number of participants | ||
|---|---|---|---|---|---|
| Paired with a cat | Returning cat before 4 months | Removed from study before 4 months | Participants remaining in study ≥ 4 months, N = 23 | Length of time before adoption, N = 22 | Adopting cat and returning cat by 12 months |
| 29 | 4a | 2b | 22 (adopting) 1 (not adopting)c | 1 month = 11 2 months = 2 3 months = 0 4 months = 9 | 1d |
Note: aTime and reason for returning the cat: (1) 2 days—did not like the smell of cat litter and did not want an indoor cat; (2) 2 months—kitten too rambunctious; (3) 2 months—participant afraid of tripping over cat; and (4) 3 months—did not like having to keep door to sewing room closed to prevent cat from getting in room.
bReason for removing two participants from study before 4 months: (1) participant fell and broke their collar bone (fall was not related to cat); (2) cat bit participant on finger while trying to trim the cat’s nails. A third participant was removed from the study at 8 months after being paired with their cat because they tripped over the cat and fell. This participant had already adopted their cat and chose to keep the cat.
cOne participant that fostered their cat for 4 months and chose not to adopt it went into the study genuinely interested in fostering a cat and helping to make it more adoptable.
dParticipant vacillated over whether to keep their cat or not. After the lockdown associated with COVID-19 ended, participant decided to return their cat.
Discussion
Our findings from this feasibility study show the potential for companionship from cats to lessen loneliness and improve mental health in older adults. These findings are noteworthy outcomes in the small number of enrolled participants and the negative impact of the COVID-19 pandemic had on some of them. Our results from the Comfort from Companion Animal Scale showed participants perceived high comfort from their cats throughout the study. A larger study population may provide the statistical power needed to show additional beneficial effects on loneliness and mental and physical health in older adults.
Our findings also showed a high interest in older adults adopting their foster cats when perceived barriers to adoption were removed. In addition, some participants benefitted from additional support from the study veterinarian during the initial months of fostering as they adjusted to living with their cats. This study also successfully placed homeless cats living in shelters into loving homes, mutually benefitting felines, foster parents, and their communities.
The present observation of a significant decrease in loneliness in older adults who gained a newly acquired companion cat is the first from a longitudinal study of this sort. An accumulating body of research has demonstrated the adverse psychological effects arising from isolation, recently including that brought about by the COVID-19 pandemic, so such a result may not be surprising (Carr et al., 2021; Emerson, 2020; Kogan et al., 2021; Phillipou et al., 2021; Ratschen et al., 2020).
The present results are consistent with some studies showing pet ownership to decrease loneliness in all ages (Kretzler et al., 2022). However, the present study differs from many previous human–animal bond studies with respect to the unanticipated COVID-19 pandemic. Because of mandated social distancing, loneliness increased in a substantial proportion of older adults during the pandemic (Mund et al., 2020). Although this study was not designed to specifically evaluate the impact of cat ownership on older adults living alone during a pandemic, some of the participants were affected by the pandemic. The pandemic effect on older adults may explain why a significant decrease in loneliness in participants was observed 4 months after placement with a cat, but it did not persist at 12 months. The data collected at 12 months were affected more by the pandemic than at baseline, 1 month, and 4 months. Other factors besides the cumulative effects of the COVID-19 pandemic that could have affected the results include the small study population and the cat presence becoming less novel over time.
In the present study, cat companionship did not show a statistically beneficial effect on physical or mental health as measured by the SF-12. However, a 4-month improvement was observed for mental health that approached statistical significance (adjusted p = .079). Also of relevance to the physical health of our subjects, risks associated with cat ownership were considered. One participant suffered a bite to a finger while attempting to trim their cat’s nails, another fell with an injury unrelated to the cat, and another tripped over their cat and fell; however, no harm resulted. Previous studies have examined the risks of injury from cat ownership in older adults, with bites, scratches, and falls emerging as leading but minor risks (Ellis & Ellis, 2014; Stevens et al., 2010). In the present study, the rates of injury were lower compared to unselected older populations, suggesting the benefits may outweigh the risks in older adults who are new cat owners. Our inclusion criteria, monitoring, and safety protocols may have contributed to our lower rates of injury.
Self-efficacy in participants remained consistent after being paired with a cat, and there appeared to be no peripandemic effect on this assessment. A study evaluated the use of this scale in 25 countries and found the mean sum score in the United States was 29.5 out of 40.0 (Scholz et al., 2002). Although the composite scores for this scale generally are not categorized as high or low self-efficacious, the mean sum scores for participants in our study were higher than the national average, ranging from 33.8 to 34.2. This suggests that participants’ ability to cope with daily challenges and to adapt after experiencing various stressful events, such as health-related problems or the impact of the pandemic, remained strong throughout the study.
The Positive and Negative Affect Scale (Mroczek & Kolarz, 1998; Watson et al., 1988) showed statistical significance was not achieved for either scale; however, similar to results from other assessments, a beneficial trend was observed for negative affect during the first 4 months of the study, that did not persist at 12 months. This suggests that participants may experience fewer negative effects after being paired with a cat, but a larger study population not affected by a pandemic is needed to determine if this outcome may reach significance.
Mean values for positive affect numerically improved at 1 month. They declined at 4 months, suggesting companionship from a cat could have more of a beneficial impact on reducing feelings of emotional distress than on increasing positive mood states. Because loneliness in older adults increases the risk for depression, the significant decrease in loneliness observed in the present study, in combination with a trend toward a reduction in feelings of emotional distress, may suggest the two conditions are related. If so, companionship from a cat may be a potential therapeutic option for older people who suffer from loneliness and depression.
The present study also demonstrated older adults who live alone with no other pets and who chose to foster a cat have a high rate (75.9%) of adopting their cat. Some participants who never had a cat before or only had an outdoor cat were comforted by the ability to foster a cat before committing to adopting one. Removing financial barriers further facilitated the ability of some participants to adopt a cat when they otherwise would not have been able to do so. Participants also favorably viewed the ability to select the cat they would foster.
About half of the participants who chose to adopt their cat decided within the first month of fostering, while others took longer. Some participants needed support during the first few months of fostering as they adjusted to living with a cat. Having the study veterinarian available to answer questions or find solutions for their concerns facilitated participants’ bonding with their cat and ultimately adopting it.
The majority of respondents did not return their cat after adoption. However, two participants passed away after completing the study, and the investigators facilitated finding the cats new homes before their deaths, which comforted both participants.
Limitations of this study included a small sample size. Although most participants completed the surveys in person, some did theirs over the phone. A control group was lacking, so it is unclear what effect the monthly home visits from investigators had on loneliness and other scales. The COVID-19 pandemic occurred part way through the study, and some participants commented that it influenced some of their responses to the survey.
Conclusion
Overall, the present results show that a cat-fostering program for older adults can be successful and beneficial both for older adults and cat adoption. Positive effects include decreased loneliness and improved mental health, and a high level of perceived comfort from a cat. Although the pandemic likely obfuscated the impact on well-being, initial results suggested that fostering a cat benefited older adults. When perceived barriers were removed, permanent cat adoption rates were high for homeless shelter cats fostered by older adults. Support from the study veterinarian also appeared important in fostering a bond between the foster cat and some participants.
Recruiting for the study was more challenging than initially anticipated, and homebound segments of the target population were even more challenging to reach because they did not see flyers about the study posted in senior centers or advertised in local publications. A story about the study that aired on a major news station generated the most interest from potential participants, but unfortunately, it aired just 2 weeks before the COVID-19 pandemic lockdown. The pandemic halted enrolling any new participants in the study for 6 months, and some potential participants adopted a pet on their own during that time. The unanticipated number of senior living facilities in the area that prohibit pets also affected recruitment. Nonetheless, the unforeseen challenges encountered during this feasibility study reinforced why such types of studies are important to conduct prior to conducting large-scale studies.
This study established important initial findings, which could catalyze more extensive and nuanced future studies. For example, this study did not explore the potential impact of previous pet ownership, and prior studies should include variables on previous pet ownership (including type of animal). Additionally, future studies should have a control group to determine the impact of cat ownership. Finally, we could not generalize our findings because of the small sample size. Therefore, future studies should be larger, with a more diverse sample of older adults, allowing for more nuanced analyses.
Funding
This work was supported by the Human–Animal Bond Research Institute (grant number HAB17-024) and Nestle Purina Pet Care (cat food and cat litter).
Conflict of Interest
S. L. Sanderson has given lectures sponsored by Nestle Purina Pet Care and has conducted prior research funded by Nestle Purina Pet Care.
Acknowledgments
We would like to acknowledge our community partners, including the Athens Area Humane Society, Campus Cats, Athens Community Council on Aging, Athens Boom Magazine, Dr. Heidi Ewen for her initial involvement and contributions to the study, and Dr. Steve Holladay for his guidance and support throughout the study.
Author Contributions
S. L. Sanderson planned the study, conducted in-person surveys, conducted monthly home visits, wrote sections and contributed to revising the paper; K. G. Emerson planned the study, conducted phone surveys, wrote sections and contributed to revising the paper; D. W. Scott conducted monthly wellness surveys, assessed participants who had adverse events, and wrote sections and contributed to revising the paper; M. Vidrine conducted monthly wellness surveys, wrote sections and contributed to revising the paper; D. L. Hartzell planned the study, conducted phone surveys, delivered cat supplies to participants and contributed to revising the paper; and D. A. Keys performed all statistical analysis.
