Pet-Assisted Therapy for Delirium and Agitation in Hospitalized Patients with Neurocognitive Impairment: A Review of Literature

Delirium leading to agitation is a common issue in elderly people and patients with underlying neurocognitive impairment. Despite use of medications to treat agitation, polypharmacy is a major concern and might lead to multiple side effects in this patient population. Therefore, it is imperative to investigate non-pharmacological methods that can provide solutions to the problem. The objective of this review was to evaluate the impact of pet-assisted therapy on elderly patients, with a major focus on agitation and delirium. For the purposes of this study, a scoping review was performed using PubMed, Google Scholar, and ClinicalTrials. We reviewed literature from 1980 to 2021. Out of the 31 studies reviewed, 14 commented on agitation with respect to pet-assisted interventions. Of these, eight studies (57%) reported a statistically significant reduction in agitation and/or delirium in patients who were exposed to pet therapy. Pet-assisted therapy can improve the standardized care in hospital-based settings for patients with neurocognitive impairment because of better companionship, reduced agitation and mood disorders, and better stability of hemodynamic status. These interventions can pave the way for better patient and hospital satisfaction.


Introduction
Delirium is an acute confusional state illustrated by altered conscious levels along with a reduced ability to focus, sustain, or shift attention. Delirium mostly develops acutely and follows a waxing and waning pattern [1]. It is usually characterized by psychomotor and autonomic hyperactivity that manifests as agitation and hallucinations [1]. Some of the important causes include severe or chronic illness, metabolic imbalances, drugs, infections, or surgery. Additionally, other risk factors include any condition that results in a prolonged hospital stay, being a resident of a nursing home, and preexisting neurocognitive disorders such as dementia.
In addition to pharmacologic interventions (antipsychotics, benzodiazepines, and cholinesterase inhibitors), many non-pharmacologic interventions are also commonly employed, which range from reducing modifiable risk factors to unorthodox methods such as aromatherapy, music intervention, massage, and multi-sensory stimulation [2]. One such nonpharmacological treatment is animal-assisted therapy (AAT), which entails the introduction of animals into patient settings in order to improve patient agitation, anxiety, and mood [3].
Very early medical texts have already described the therapeutic effects of animal companionship. Hippocrates and Galen were early proponents, suggesting horse-riding as a cure to insomnia and to prevent disease [4]. The structured use of animals as an aid in treating mental and physical health disorders dates back to 1792 [4].
Indications for the use of animal-assisted therapy in the literature vary widely, but mostly focus on diseases with neurological or psychiatric components [3]. The elderly population has been especially studied in preference to the pediatric and adult population [3]. The outcomes of the studies, while generally positive, have proven difficult to quantify as a result of vastly different patient settings, length of animal encounters parameters measured, and the indication for the initiation of AAT. While most of these therapies have been usually studied in the setting of dementia, outcomes suggest they may also have a place in treating acute confusional states and agitation.
The objective of this review was to evaluate the impact of pet-assisted therapy on elderly patients, with influence on agitation and delirium as the primary outcome. The secondary outcome highlighted cognition, quality of life, physical functioning, vitals, and depressive/mood disorders.

Methods
For the purposes of this study, a scoping review was performed. PubMed, Google Scholar, and ClinicalTrials.gov were used for the retrieval of studies required for the review. We reviewed literature from 1980 to 2021. The keywords used in the review were "dementia", "agitation", "delirium", "pet assisted therapy", "cognition", "elderly", and "psychiatry".

Studies Reviewed for Primary and Secondary Outcomes
This scoping review included 31 articles that studied the impact of pet-assisted therapy on the health of elderly patients, including 16 interventional studies, 9 randomized controlled trials, 4 observational studies, and 2 studies with a nonequivalent control group pretest−posttest design. The primary outcome analyzed in this review was the effect of pets or pet robots on agitation or delirium. The secondary outcome analyzed in this review was the effect of pets or pet robots on cognition, depression and loneliness, quality of life, blood pressure, physical functioning, and activities of daily living (ADL).

Population Studied
The population of interest in this study is older adults, especially those with dementia. The studies analyzed in this review all examined the effect of pet therapy on older patients, but the inclusion criteria for this population varied among studies. Of these studies, 15 utilized populations of older patients residing at long-term care facilities, 11 focused on the population of elderly adults in nursing homes, and 3 focused on elderly patients in inpatient units. Eight studies used an age criterion of >65, two studies used an age criteria of >60, one study used an age criteria of >55, and one study used an age criteria of 80 to 90 years of age. A further 24 studies narrowed their study populations by only including patients with dementia.

Intervention
Of the 31 studies included in this analysis, 12 (39%) studied living pets as a form of therapy. The majority of studies utilized dogs or cats, but two studies also included fish, and one included rabbits, birds, and horses as therapy animals as well. In contrast, 14 studies (45%) investigated pet robots as a form of therapy. Interventions ranged from individual therapy sessions to group therapy session to cohabitating with an animal.

Review
Summarized data from all articles analyzed in this review are reported in Tables 1 and 2 .

Article Primary Outcome Secondary Outcomes Limitations
Majic et al. [5] Reduction in agitation in AAT group (p < 0.05) Reduction in depression in AAT group (p < 0.05) • Small follow-up period Nordgren and Engström [6] Physical non-aggressive behaviors decreased over the period, but this was non-significant (p > 0.05) The Physical aggressive behaviors were reduced in the intervention group, but the decrease was non-significant (p > 0.05) Overall facial expressions improved (p > 0.05) but more happiness was observed in participants receiving the intervention (p < 0.05); significant social interactions with the intervention included talking with the activity coordinator and staff (p < 0.05) • Small sample size Olsen et al. [8] No significant effects on agitation (p > 0.05).
No significant effects of the intervention were found from T0 to T1 for depression (p > 0.05) • Not specified Nurenberg et al. [9] Agitation scores were significantly reduced in cases of points less than 2 (p < 0.05); violent incidences were significantly reduced (p < 0.05) Improved intrusiveness was associated with reduced violence (F = 5.62, df = 1 and 76, p = 0.02) and with a diminished group effect (F = 1.91, df = 3 and 76, ns) • The groups had been divided into many subgroups given on the type of therapy received which reduced the number of participants per group. Churchill et al. [10] Reduced agitation (p < 0.05) Not specified • Not specified Richeson et al. [11] The agitated behaviors of the participants decreased immediately following the intervention phase and increased during the follow-up phase of the pilot study Social interactions increased significantly from the first week to the last week of the AAT intervention • Small sample size Jøranson et al. [12] Reduction in agitation in PARO versus TAU at 3-month follow-up (p < 0.05) -Reduction in depression in Paro versus TAU at 3-month follow-up (p < 0.05) -In those with severe dementia, quality of life scores did not decrease in the PARO group • Patients were not blinded Libin and Cohen-Mansfield [13] Physical agitation and overall agitation decreased with the plush cat (p = 0.046, respectively); interactions with the robotic cat also lowered the level of agitation, but it was not significant Significant increase in pleasure (p < 0.01) and interest (p < 0.05) scores while playing with plush cat • Not blinded, small sample size Sellers et al. [14] The results indicated a statistically reliable decrease for the total agitated behaviors category (t = 7.05, p < 0.0001) Improvement in social behaviors (p < 0.05) • Small sample size

Article Primary Outcome Secondary Outcomes Limitations
Moyle et al. [15] PARO was more effective than the usual care for improving agitation (p < 0.05) • Participants in the PARO group were more verbally engaged than participants in the plush toy group (p < 0.05) • PARO was more effective in improving pleasure (1.12, 95% CI: 1.94-0.29, p = 0.008); videos showed that when measured using CMAI-SF, there was no difference between groups • Small duration for intervention, missing data protocol Friedmann et al. [16] Agitation decreased in the PAL group (p = 0.423) and remained the same in the reminiscing group (p = 0.865) •

Article Primary Outcome Secondary Outcomes Limitations
Moretti et al. [21] Agitation/delirium not investigated Baek et al. [25] Agitation/delirium not investigated • No significant differences observed in terms of depression or problematic behaviors • Cognition improved in AAT group; more improvement was seen at week 8 than at week 4 (p < 0.001) • Mood scores decreased in the AAT group; more improvement was seen at week 8 than week 4 (p < 0.014) • Small sample size • Post hoc analysis points were not specific • Response bias; inaccuracy on account of elderly people responding • A significant decrease in the MMSE score in the NAO group, delusions (a significant increase in the NAO group), apathy (a significant decrease in the NAO group), and irritability/lability (a significant increase in the PARO group • Many patients lost to follow-up Lane et al. [31] Not discussed • Statistically significant increases in positive patient behavioral states when comparing the same group (p < 0.05) • Presence of pre-PARO positive behaviors was significantly associated with post-PARO positive behaviors (p < 0.05); the presence of pre-PARO negative behaviors was significantly associated with post-PARO negative behaviors (p < 0.05) • Decreases in negative patient behavioral states (p < 0.05) • Small sample size Kramer et al. [32] Not discussed • All patients exhibited various degrees of interactive behavior • Small sample size Šabanović et al. [33] Not discussed • PARO increases specific social interactions and activity levels • Small sample size Chu et al. [34] Not discussed • Increased engagement in the majority of participants • Not specified Kongeable et al. [35] Not discussed • The presence of the dog increased the number of total social behaviors of the AD clients, but no differences were found in behaviors between its temporary and permanent placement • Not specified

Primary Outcomes
Fourteen out of thirty-one studies in this analysis investigated the primary outcomes. Of these, eight studies (57%) reported a statistically significant reduction in agitation and/or delirium in patients who were exposed to pet therapy, and six (36%) reported no statistically significant difference. The majority of the studies included in the review reported decreased agitation and aggression in patients [5][6][7][8][9][10][11][12][13][14][15][16][17][18]. There are certain characteristics of the animalassisted interventions that might be responsible for the decrease in agitation. The pets, and their actions and voices, overall, have a calming influence over the subjects and might even help in emotional expression in these cases [36]. These interventions remind patients of the comfortable environment at home [36]. Patients could also have a false sense of distraction from their current environment, which could assist the healthcare team in providing optimal care [36].
With regard to pet assisted therapy, dogs were the most frequently employed animals given their training potential and social nature [37]. However, numerous other animals, including robotic pets, have also been utilized. The comfort level might be more pronounced if the patient's preference is taken into account and the choice of pets is not restricted [36]. The positive impact might also be increased in patients who are not visited by family members at all, either due to time constraints or geographical limitations [36,37]. The result is that an equally important facet of patient care in this population, namely companionship, begins to develop, and there is less sense of being dependent on an individual [37].
There are also rare cases where agitation is increased instead of a desirable effect [38]. This could possibly be explained by the idea of hyperactive delirium that might have been present in these patients. A paper published in 1985 presented a possible theory. Hyperactive delirium, seen postoperatively, is associated with levels of beta-endorphin and cortisol [39]. Endorphins are naturally increased when a pleasing stimulus, for example, a pet, comes into view. Therefore, this cycle of a negative impact could be initiated [39].
Two systematic reviews and one meta-analysis also discussed the impact of animalassisted intervention on agitation [40][41][42]. However, the reviews were focused on a few articles and excluded a number of articles with important findings pertaining to agitation. Hughes et al. included Majić et al., Richeson, Sellers et al., and Nordgren and Engström, but the review did not include other important articles and did not summarize a focus on agitation as done for other aspects [5,6,11,14,40]. Similarly, Park et al. did not include Sellers et al., Richeson, or Nordgren and Engström, who noted a significant reduction and exacerbation in agitative behaviors in these patients [6,11,14,41]. A similar observation was made by Bert et al. [42]. Therefore, the findings of these reviews have to be validated in a larger setting with an appropriate intervention.

Secondary Outcomes
An improvement in cognition was noted in 22% of the articles reviewed. A significant decrease in depression was also reported by Friedmann et al. in the PAL group (p = 0.013) and by Abdi et al. in 11/33 studies [16,43]. Furthermore, owing to reduced mood disorders, cognition in these patients can improve as observed by a few studies [5,12,14,23,27,29]. Improvement in loneliness were also reported by 16% of the studies. Animal-assisted interventions were also responsible for a reduction in anxiety, sadness, and irritability [27,29]. The possible explanation for this could be that animal-based interactions might ease suffering and build neuronal networks not targeted by pharmacological methods [44]. Furthermore, because these patients usually live in a very lonely environment, pets could help in potentially humanizing the wards to debunk some of the negative opinions that have arisen over the years [44]. Once patients engage more in the rehabilitating efforts done by the healthcare teams as a result of these interventions, quality of life might be improved, as reported by Bert et al. [42].
Animal assisted interventions might also improve physical functioning and activities of daily living, as witnessed by Friedmann et al. and Cherniack and Cherniack in the PAL group compared with the reminiscing group (p = 0.306 vs. 0.072) [16,45]. This might be due to patient curiosity about the actions and interactions of the pets [36]. Other reasons also include distraction from chronic pain and a sense of responsibility for the pets in question [46].
The improvement in physical functioning is not only limited to depression and cognition; the overall health of the patients might improve as well, for example, a reduction in systolic blood pressure was reported by a few articles [16,19]. All these changes can improve the overall quality of life and decrease the length of stay in hospitals [16].

Pet-Assisted Interventions in Hospital Setting
Hospitals are high-risk areas, thus introducing animals has to be carefully considered.
There are a few risks associated with the intervention, such as infections, allergies and animal accidents [47][48][49][50]. The odds of contracting zoonosis and MRSA are higher [51,52]. Additionally, the more chronic patients would require these interventions the most. Patients with chronic illnesses include immunocompromised or malnourished categories. However, these categories were excluded from a majority of the reviews due to safety concerns. Therefore, there is a need to personalize these interventions in a way that benefits all categories. Perhaps the answer to this issue lies in using technology to invent robotic pets for this subset of patients.
Polypharmacy is a concern in this patient population, and conventional therapies that include drugs such as antipsychotics and benzodiazepines that have considerable side effects. Patients with cognitive impairment also frequently require assistance at home especially in regards to drug administration. These factors together add considerable cost on top of the original patient treatment, and is an area where pet assisted therapy may help. Thus, this is another area where pet-assisted therapy may be especially useful because the approach could be used to reduce resources utilized in a hospital setting in arranging home care, the administration of medication, and overall patient load so that optimal healthcare can be ensured. Additionally, there would be minimal risk of relapse because less medications would be used, reducing the need for longer stays to taper medications. The length of stay of most patients with underlying neurocognitive impairment would decrease in a hospital-based setting should pet assisted interventions be implemented because of a sense of safety and companionship. This might help the patients in engaging in rehabilitation regimens better than before. The quality of stay would also improve in a hospital setting because less patients would have hypertensive disorders owing to abrupt changes in their hemodynamic status. Overall, the satisfaction rate in the hospital would increase, not just for the patients but also for the healthcare providers, because the sense of responsibility of helping patients to have a better life out of the hospital is somewhat fulfilled.
There were a few limitations to the review. Some of the articles included were small centered studies, and the limited data available regarding this field makes the generalization of results a dilemma. In many articles, the search was limited to English languagebased databases provided a very narrow ground for exploring more cases. Inclusion and exclusion criteria were not clearly stated in a few cases, necessitating the need for assumption, which might have introduced bias. In a few cases, the data were very heterogeneous and made the comparison of data an uphill task. Lastly, some articles did not control confounding variables such as exercise, which could have had some influence on the results.

Conclusions
Pet-assisted therapy can improve the standardized care in hospital-based settings for patients with neurocognitive disorders because of better companionship, reduced agitation and mood disorders, and better stability of the hemodynamic status. These interventions can pave the way for better patient and hospital satisfaction. However, to truly evaluate the novel invention, randomized controlled trials should target these methods in a hospitalbased environment. Institutional Review Board Statement: The study was conducted in accordance with principles of Helsinki.