Effects of dog-assisted intervention on quality of life in nursing home residents with dementia

Abstract

Background: People with dementia often have a poor quality of life. Therefore, methods that can improve their life situation must be identified. One promising method is dog-assisted intervention.

Aim: This study aimed to investigate the effects of dog-assisted intervention on quality of life in nursing home residents with dementia.

Materials and methods: A one-group, pretest post-test study design was used. Quality of life was measured using the QUALID in 59 nursing home residents prior to and after a dog-assisted intervention. Non-parametric tests were used to analyze the data, and effect sizes were calculated.

Results: The participants’ total scores improved significantly between baseline and post-test 1 (p = < 0.001) and worsened significantly at post-test 2 (p = 0.025). The largest effect size was found for the item ‘Verbalization suggests discomfort’ (p = 0.001).

Conclusion: The results indicate that dog-assisted interventions can have positive effects on quality of life in nursing home residents with moderate to severe dementia.

Significance: The results contribute to a growing knowledge base about non-pharmacological methods that can be used in dementia care. Occupational therapists should consider dog-assisted interventions when planning activities that can reduce the illness burden and improve the quality of life for people with dementia.

Keywords:

• Animal-assisted therapy
• dementia
• occupational therapy
• quality of life

Background

Dog-assisted interventions (DAIs) are interventions that use dogs to treat humans for specific purposes. Three major directions for DAIs have been identified: therapeutic purposes, activating purposes, or as support for specific procedures [1]. In Sweden, therapy dogs, or rather, therapy dog teams (i.e. a dog handler and a dog), are trained and certified in accordance with a national standard [2]. Therapy dog teams work in different contexts. The dogs are used with positive results in healthcare settings in relation to people with dementia or psychiatric disorders [3,4]. A growing scientific evidence base indicates that DAIs can reduce behavioral and psychological symptoms and/or improve quality of life (QoL) in people with dementia [5–11]. Moreover, the use of therapy dogs can possibly slow down the progression of cognitive impairment in patients with early stage or mild Alzheimer’s disease [12].

In Sweden, a therapy dog, or a therapy dog team, works goal-oriented, and the intervention is planned and documented in the patient’s or resident’s medical chart. Commonly, DAI is prescribed for a certain indication by a professional such as an occupational therapist, a physiotherapist or a registered nurse [13]. However, despite growing knowledge within the research field, the evidence base is preliminary and rather weak [1,5], mainly because of small studies with small sample sizes that make it difficult to generalize the findings. The aim of the present study was to investigate the effects of DAIs on QoL in nursing home residents with moderate to severe dementia. Another aim was to investigate whether any differences concerning the effects on QoL are related to gender, age, or type of dementia.

Materials and methods

Setting

The present study was conducted in three different nursing homes located in a Swedish medium-sized municipality. At these nursing homes, dog-assisted interventions are in use since 2011 and are thus considered routine. All residents at the nursing homes are eligible for receiving DAIs on prescription by an occupational therapist, a physiotherapist, or a registered nurse. However, people with dementia are prioritized.

At monthly interprofessional team meetings (i.e. meetings between nursing staff, occupational therapists, section managers, physiotherapists, and registered nurses), the residents’ situations and needs for interventions that can improve the residents’ well-being are discussed. One such intervention is DAI. Common indications for DAI are that a resident is increasingly anxious or depressed or that the nursing staff perceives that the intervention would improve the resident’s QoL. Based on each resident’s problems (mostly behavioral or psychological symptoms), the team discusses the aim and goal of the intervention. The problem, aim, goal, and content of the intervention; how the intervention will be evaluated; and how the intervention will be followed up is documented in the resident’s care plan.

In the present nursing homes dog-assisted intervention usually consists of ten one-on-one sessions twice a week, but sometimes only once a week. During each session, the therapy dog team and the resident perform activities in accordance with the prescription. The choice of activity is based on the resident’s current health status and needs. The activity can be performed in the resident’s flat or during an indoor or outdoor walk. Thus, different activities can be performed such as playing with the dog, walking the dog, grooming, or sitting down or lying on the bed petting and talking with the dog. A session usually lasts for 20–45 min depending on what the resident manages for the time being.

The nursing staff assesses the resident’s current QoL prior to the intervention. The assessment is repeated twice after the completed intervention, both immediately after and 1 month after. For the assessment, the nursing staff uses the Quality of Life in Late-Stage Dementia scale (QUALID). The results of the assessment are registered in the residents’ medical chart and used for evaluation by the nursing staff. See Box 1 for an authentic description of DAI.

Box 1 Description of a dog-assisted intervention. The description is based on an authentic registration in a resident’s medical chart. However, the text has been translated into English by the authors.

Bob and Jack

Bob was a 75-year-old nursing home resident originating from another country. He had a medical history of severe vascular dementia, status post stroke, atrial fibrillation, previous alcohol abuse, rectal bleeding, alcoholic encephalopathy, cerebral degeneration, gastro-oesophageal reflux disease with oesophagitis, and diaphragmatic hernia. The nursing staff had observed that Bob was getting increasingly anxious and showed signs of discomfort. He repeatedly asked for his legal guardian and was physically inactive. At a team meeting, it was decided that Bob should receive 10 sessions of DAI twice a week.

The dog handler’s assignment included a goal for the intervention, that is, Bob needed to be more physically active, and he needed to find something else to focus on instead of asking for his legal guardian all the time. Prior to the intervention, Bob’s contact nurse completed the QUALID questionnaire. Bob’s total score was 36. After the intervention, Bob’s total score was 25, indicating improved quality of life. One month later, Bob’s score was once again 25. The greatest improvement was observed for verbal signs of discomfort. Prior to the intervention, verbal signs of discomfort were observed more than once each day (i.e. Bob was repeatedly asking for his legal guardian). After DAI, verbal signs of discomfort occurred rarely or never. One month later, verbal signs of discomfort were observed less than once a day.

The dog handler documented the intervention in Bob’s care plan:

Visit [V] 1: Today, Bob had his first visit by Jack, the therapy dog. Bob joined us for an outdoor walk; he used his walker and held Jack’s leash. Bob was verbally active during the whole visit and reached and touched Jack spontaneously. At the end of the visit, he would talk straight to Jack. Bob also fed Jack dog treats.

V2: Today, Bob and I went for an outdoor walk with Jack. Bob used his walker. He was verbally active and reached spontaneously out for Jack’s back. During the visit, Bob made jokes and laughed. He mentioned his legal guardian twice. He speaks his native language with me [authors’ note: the dog handler has the same country of origin as Bob].

V3: Today, Bob took Jack for an indoor walk. Bob was verbally active, talked straight to Jack, and touched him. In the library, Bob played board games with Jack. He managed to follow my instructions to fill the game board. Bob was cheerful and laughed many times. He mentioned his legal guardian twice.

V4: Today, Bob went for an outdoor walk with Jack. He used his walker and held Jack’s leash. He stopped to pat Jack several times. Before we finished, he fed Jack treats while searching for eye contact, and he would talk directly to Jack.

V5: Today, Bob and Jack went for an indoor walk. Right away, Bob initiated verbal contact with Jack, reached spontaneously out to touch Jack several times, and talked straight to him. Bob played games with Jack in the library and managed to fill the game board in accordance with my instructions. Bob cheerfully told me old memories.

V6: Today, Bob followed Jack and me for a short outdoor walk. Bob used his walker. He was verbally active and touched Jack over the whole body and would talk directly to Jack.

V7: Today, Bob went for a short walk with Jack. Bob held the leash, stopped several times, touched, and talked to Jack. Bob was feeding Jack treats, he was verbally active, laughed, and made jokes.

V8: Today, Bob did not want to see Jack.

V9: Today, Bob went for an outdoor walk with Jack. Bob was verbally active during the whole visit and reached out several times to touch Jack, and he would talk directly to Jack.

V10: Today, Bob walked to the library with Jack. He played games with dog chews and managed to follow my instructions to fill the game board. Bob smiled on several occasions, talked directly to Jack, and patted Jack spontaneously. When we returned to his flat he asked once for his legal guardian. This was the final session. The QUALID assessment has been filled.

Design

A one-group, pretest – post-test study design was applied. A retrospective study was conducted based on the nursing staffs’ assessments of the residents’ QoL prior to and twice after the DAI. Post-test 1 was conducted one to 2 weeks after the intervention, and post-test 2 was conducted 1 month later.

Sample

All residents with moderate or severe dementia in 3 municipal nursing homes in a medium-sized town in the middle of Sweden who had DAI initialized during January 2014 to December 2016 (n = 66) were included.

Instrument

The Quality of Life in Late-Stage Dementia scale (QUALID) was developed by Weiner and coworkers in 2000 [14]. It is a questionnaire specifically developed to measure QoL in later stages of dementia. The assessment is made by proxy informants. In the present study, nursing staff familiar with each resident made the assessments. The questionnaire consists of 11 statements about positive and negative behaviors indicative of QoL in late stages of dementia [15]. The responses are graded on occurrence during the last week on a five-point scale. The points are summed up to a grand total between 11 (best QoL) and 55 (worst QoL). The questionnaire is administered in a structured interview format and takes approximately 5 min to complete. The proxy informants must be familiar with the person’s normal behavior. The Swedish version of the QUALID questionnaire has been found to have a satisfactory level of internal consistency (Cronbach’s α = 0.74), good test-retest reliability (ICC = 0.86), and satisfactory inter-rater reliability (ICC = 0.69). It has also shown good criterion validity, good clinical validity, and moderate responsiveness to change (SRM = 0.66) [15].

Procedures

The first author (an occupational therapist) identified residents who had received DAI during the specified time at the three nursing homes. The QUALID assessments were copied from the residents’ medical charts. The first author also used the medical charts to collect data on the residents’ age, sex, and medical diagnoses. Prior to handing data over to the second author, the data were anonymized.

Data analysis

The responses were coded, entered, and verified using the Statistical Package for Social Sciences (SPSS) version 22 [16]. In order to be able to compare results, scores from residents who had received more than ten sessions of the intervention (n = 7) were excluded from analysis. Demographic differences were compared with the chi-square test [x²] for categorical variables (gender, diagnosis, indication) and with t-tests for continuous variables (age).

Next, age, diagnosis, and indication were coded into sub-categories. Age was dichotomized based on the median into a ‘61- to 82-year-old’ group (46.4%) and an ‘83- to 93-year-old’ group (53.6%). Diagnoses was coded as follows: 1 = Alzheimer’s disease, 2 = dementia NS, 3 = vascular dementia, and 4 = other dementia. There were several indications for the intervention, and some participants had more than one indication. To facilitate the analysis, the different indications were coded. The sub-categories were as follows: 1 = ‘Reduce anxiety’ (e.g. to reduce low mood, worries or stress and pain relief); 2 = ‘Sustain/improve QoL’ (e.g. to perceive wellbeing, joy or safety); 3 = ‘Other indication’ (e.g. socialization, reduce isolation, adaption, improve/sustain physical capacity or functions for activities in daily life and more); and 4 = ‘1 + 2’ (both reduce anxiety and sustain/improve QoL).

For total scores and for each questionnaire item, analyses of the differences between and within categories were conducted. At each time point, the participants’ total scores were compared by the Related-samples Sign Test. For differences between and within categories, non-parametric tests were used. For differences between categories (i.e. independent samples), the Mann-Whitney U test (gender and age) and the Kruskal-Wallis test (diagnosis and indication) were employed. To follow-up significant results from the Kruskal-Wallis tests, pairwise comparisons with adjusted p-values were conducted [17]. For differences within categories (i.e. related samples), the Wilcoxon Signed Ranks Test was used.

In accordance with Field [17], effect sizes were calculated. Effect sizes (r) <0.30 are considered small and effect sizes greater than 0.50 are considered large. The significance level was set at < 0.05.

Ethical considerations

The study protocol was approved by the regional ethical vetting board in Stockholm (Dnr 2015/2264-31/5).

Results

In total, 190 QUALID assessments concerning 66 residents were identified. Incomplete assessments were excluded. Consequently, at baseline results from 59 individuals were available (Table 1). At post-test 1, complete QUALID assessments were available for 58 participants, and at post-test 2, data were available for 44 participants. One person passed away during the study time and data were missing for 13 participants (i.e. no documentation of the QUALID assessment, incomplete forms, or missing dates).

Table 1. Baseline characteristics of the participants in the study (n = 59).

Baseline characteristics
Variable	Frequency (%)
Gender
Female	34 (57.6)
Male	25 (42.4)
Age
Mean (SD)	81.0 (7.2)
Range	61 – 93
Diagnosis
Alzheimer’s disease	14 (23.7)
Vascular dementia	12 (20.3)
Frontotemporal dementia	1 (1.7)
Dementia NS	24 (40.7)
Mixed dementia	6 (10.2)
Alcohol-related dementia	2 (3.4)
Indication for intervention
Reduce anxiety	22 (37.3)
Sustained or improved QoL	21 (35.6)
Both reduce anxiety and sustain/improve QoL	6 (10.2)
Other reason	10 (16.9)

The majority of the participants were women with an average age of 83.7 years (Table 1). Their functional capacity varied. Some participants were able to walk without any walking aid while some used a walker or a wheelchair. Some participants were sitting on a chair when having contact with the dog while some were lying on the bed or resting on the couch. It could also be different at different occasions. Non-specified dementia was the most common dementia diagnosis and most participants received the intervention in order to reduce anxiety.

The same dog handler with two different therapy dogs conducted all the interventions. The dog handler was an experienced assistant nurse who had been working with DAIs since 2011. Both dogs were experienced and certified flat-coated retrievers. The intervention included 7 to 10 appointments with the therapy dog team.

The baseline tests (i.e. pretest) were conducted prior to the intervention. Post-test 1 was completed soon after each participant’s final appointment with the therapy dog team. On average, the mean time between baseline and post-test 1 was 58 days (Table 2). Post-test 2 was conducted 1 month after post-test 1 (Mean [M] = 30 days) and the average time between baseline and post-test 2 was 88 days (Table 2).

Table 2. 'Significant differences in QUALID scores across time.

Total scores at different time points^a
	Range	M	SD	Mdn
Baseline	14–41	24.6	7.7	24.0
Post-test 1	11–41	20.7	6.7	19.5
Post-test 2	12–44	23.1	7.6	21.0
Differences in total scores between time points for sub-categories^b
Baseline – post-test 1
Sub-category	Mdn	T	p	r
Female	25–20	93	0.002	−0.40
Male	20–18	29	0.003	−0.40
Age 61 to 82	22–18	34	<0.001	−0.44
Age 83 to 93	24–20	88	0.009	−0.35
Alzheimer’s disease	26–21	21	0.05	−0.26
Vascular dementia	24–20	0	0.008	−0.35
Other dementia	24–18	1	0.01	−0.34
Sustain/improve QoL	21–18	25	0.003	−0.40
Reduce anxiety	28–23	31	0.02	−0.31
‘1 + 2’	28–22	0	0.04	−0.27
Post-test 1 – post-test 2
Female	20–23	255	0.01	0.38
Age 83 to 93	20–23	204	0.01	0.38
Alzheimer’s disease	21–26	47	0.05	0.30
Vascular dementia	20–22	27	0.03	0.33
Sustain/increase QoL	18–20	101	0.02	0.36
Number of days between time points
	Range	M	SD	Mdn
Baseline - post-test 1	29–141	58.0	26.0	50.5
Post-test 1 - post-test 2	15–40	30.0	4.7	31.0
Baseline – post-test 2	59–163	88.0	25.7	80.0

^a Lower scores indicate better quality of life.

^b Negative effect sizes indicate higher scores at the first time point compared to the second time point, and vice versa for positive effect sizes.

Overall, the total QUALID scores improved significantly at post-test 1 compared to baseline (p = < 0.001) (Table 2). At post-test 2, the scores again declined significantly (p = 0.025). No significant differences in total scores were observed at each time point (i.e. baseline, post-test 1, post-test 2) for gender, age or diagnosis. For indications, a Kruskal-Wallis test (H) indicated possible differences in the total scores at baseline (H(3) = 12.1, p = 0.007) and at post-test 1 (H(3) = 9.50, p = 0.023). However, analyses of subsequent pairwise comparisons with adjusted p-values did not identify any significant differences.

Table 2 shows significant differences within sub-categories between different time points. Between baseline and post-test 1, the largest effect sizes (indicating significant improvements at post-test 1 compared to baseline) were found within following sub-categories; women, men, participants aged 61–82 years, and participants who received DAI with the purpose of sustained or improved QoL. Between post-test 1 and post-test 2, the largest effect sizes (indicating a significant decrease between post-test 2 and post-test 1) were found within the sub-categories women and for participants aged 83–93 years.

Analyses were conducted for each questionnaire item. Table 3 shows descriptive statistics for each item at each time point (i.e. baseline, post-test 1, post-test 2). Table 3 also shows differences in the median scores between the different time points (baseline vs. post-test 1, post-test 1 vs. post-test 2). Between baseline and post-test 1, significant differences were found for the items ‘Smiles’, ‘Facial expression of discomfort’, ‘Appears physically uncomfortable’, and ‘Verbalization suggests discomfort’. The largest effect size (indicating a significant improvement at post-test 1 compared to baseline) was found for the item ‘Verbalization suggests discomfort’.

Table 3. Individual QUALID item scores across time.

	Descriptive statistics for questionnaire items^a^,^b
	Baseline n = 59			Post-test 1 n = 58			Differences Baseline - post-test1		Post-test 2 n = 44			Differences Post-test1 - post-test2
Questionnaire item	M	Mdn	SD	M	Mdn	SD	p	r	M	Mdn	SD	p	r
Smiles	1.95	1.0	1.14	1.56	1.0	0.96	0.01	−0.32	2.00	2.0	1.11	0.01	0.37
Appears sad	2.27	1.0	1.57	2.05	1.0	1.57	0.57	−0.07	2.24	1.0	1.71	0.63	0.07
Cries	1.58	1.0	1.19	1.28	1.0	0.91	0.06	−0.24	1.29	1.0	0.87	1.00	0.00
Facial expression of discomfort	2.78	3.0	1.33	2.19	2.0	1.07	0.005	−0.37	2.58	3.0	1.50	0.19	0.21
Appears physically uncomfortable	1.66	1.0	1.25	1.29	1.0	0.82	0.01	−0.32	1.62	1.0	1.28	0.02	0.33
Verbalization suggests discomfort	3.07	3.0	1.76	2.05	1.0	1.52	0.001	−0.46	2.69	2.0	1.73	0.02	0.35
Is irritable and aggressive	2.29	2.0	1.62	1.98	2.0	1.25	0.44	−0.10	2.04	1.0	1.35	0.82	0.12
Enjoys eating	1.37	1.0	0.96	1.37	1.0	0.94	0.61	0.08	1.53	1.0	1.16	0.51	0.10
Enjoys touching/ being touched	2.42	2.0	1.09	2.24	2.0	1.17	0.28	−0.14	2.09	2.0	1.06	1.00	0.00
Enjoys interacting with others	2.44	2.0	1.21	2.39	2.0	1.20	0.60	−0.07	2.51	2.0	1.27	0.29	0.16
Appears calm and comfortable	2.55	2.0	1.39	2.28	2.0	1.27	0.40	−0.11	2.56	2.0	1.37	1.00	0.00

^a Lower scores indicate better quality of life.

^b Negative effect sizes indicate higher scores at the first time point compared to the second time point, and vice versa for positive effect sizes.

Between post-test 1 and post-test 2, significant differences in median scores were found for the items ‘Smiles’, ‘Appears physically uncomfortable’, and ‘Verbalization suggests discomfort’. The largest effect size (indicating a significant decrease between post-test 2 and post-test 1) was found for the item ‘Smiles’ (Table 3).

Table 4 shows significant differences within sub-categories between different time points. Between baseline and post-test 1, the largest effect sizes (indicating significant improvement at post-test 1 compared to baseline) were found for women and participants aged 61–82 years concerning the item ‘Verbalization suggests discomfort’. Between post-test 1 and post-test 2, the largest effect sizes (indicating a significant decrease between post-test 2 and post-test 1) were found for women concerning the items ‘Smiles’ and ‘Verbalization suggests discomfort’.

Table 4. Significant results for item-by-item analyses of differences within Sub-categories between time points (baseline vs. post-test 1; post-test 1 vs. post-test 2).

Differences between time points within sub-categories
Questionnaire item Sub-category	Mdn1 – Mdn2	T	p	r
Baseline – posttest 1^a^,^b
Smiles
Male	2–1	18	0.05	−0.26
Age 61 to 82	2–1	18	0.05	−0.26
Facial expression of discomfort
Female	3–2	30	0.01	−0.33
Age 61 to 82	3–2	17	0.008	−0.35
Vascular dementia	3–2	3	0.03	−0.28
Reduce anxiety	3–2	10	0.02	−0.28
Sustain/improve QoL	3–2	7	0.03	−0.28
Appears physically uncomfortable
Male	1–1	0	0.01	−0.32
Age 61 to 82	1–1	2	0.01	−0.32
Verbalization suggests discomfort
Female	4–2	22	0.002	−0.41
Male	2–1	5	0.02	−0.30
Age 61 to 82	3–1	3	0.002	−0.41
Age 83 to 93	3–2	22	0.02	−0.32
Alzheimer’s disease	4–2	2	0.02	−0.31
Dementia NS	2–1	6	0.03	−0.28
Other dementia	5–1	1	0.04	−0.26
Reduce anxiety	4–2	6	0.03	−0.29
Sustain/improve QoL	5–2	12	0.009	−0.34
Posttest 1 – posttest 2
Smiles
Female	1–2	63	0.005	0.42
Alzheimer’s disease	1–3	27	0.02	0.34
Vascular dementia	1–2	15	0.04	0.31
Reduce anxiety	1–2	26	0.03	0.36
Facial expression of discomfort
Female	2–3	165	0.02	0.35
Alzheimer’s disease	2–4	46	0.05	0.30
Appears physically uncomfortable
Female	1–1	15	0.03	0.32
Alzheimer’s disease	1–1	15	0.04	0.31
Verbalization suggests discomfort
Female	2–4	133	0.007	0.41
Enjoys touching/being touched
Male	2–3	15	0.04	0.31

^a Lower scores indicate better quality of life.

^b Negative effect sizes indicate higher scores at the first time point compared to the second, and vice versa for positive effect sizes.

Discussion

The results showed that the total QUALID score improved significantly from baseline to post-test 1, indicating that DAI can have positive effects on QoL in people with moderate to severe dementia. This result is in line with other studies [1]. For example, Olsen, Pedersen [9] found significant effects of DAI on depression and QoL for nursing home participants with severe dementia. In the present study, significant differences in relation to signs of sadness and discomfort were observed, indicating that these symptoms decreased between baseline and post-test 1. These results are important from a clinical point of view. Verbal and physical signs of sadness and discomfort are recognized as symptoms of agitation. If agitation is reduced, the person with dementia will become more satisfied and relaxed. Often this means they are calmer, more communicative, and more positive. It has previously been suggested that DAI addresses unmet needs in people with dementia by providing meaningful activity, stimulation, social interaction, and comfort [9,18]. Therefore, the present findings suggest that DAI can be helpful to reduce symptoms of agitation in nursing home residents with moderate to severe dementia.

For relatives and nursing staff, communication with people with dementia can be challenging, and it is helpful to have a subject to talk about that engages the person with dementia. Discussing the dog can help the patient open up by triggering positive memories or pleasant feelings [11]. Communication interventions and an optimistic atmosphere can have a positive impact on QoL and can also assist nursing staff in different caring situations [19] and social interactions [9,11,20–23].

The present results indicate a positive relationship between DAI and indications of QoL in people with moderate to severe dementia. In a recent study about animal-assisted activity in home-dwelling people with dementia or cognitive deficits, Olsen, Pedersen [24] used the QUALID scale to assess QoL. However, they did not observe any significant effects on QoL. There are two major differences in the study by Olsen, Pedersen compared to the present study. First, Olsen and coworkers used a group-based intervention, i.e. animal-assisted activity, which is not the same as DAI. In the present study, each DAI was planned, individually tailored, and conducted from a person-centred perspective. Second, the participants in the study by Olsen and coworkers were home-dwelling persons with dementia, and only one of the participants had severe dementia. All participants in the present study were residents in different nursing homes, and all had moderate to severe dementia. Notably, however, the participants’ cognitive levels were not measured in the present study. According to the present results and based on long and proven experience, DAI may be able to sustain or improve QoL in people with moderate to severe dementia.

The present findings showed significant improvements in total scores between baseline and post-test 1 for all sub-categories except for ‘dementia NS’ and for those who received DAI for ‘other indication’. The largest effect sizes were found for women, men, people aged 61–82 years, and for those who received DAI with the purpose of sustained or improved QoL. However, since the total scores increased once again at post-test 2, no conclusions can be drawn about the remaining long-term effects of DAI. Majic, Gutzmann [25] studied the mid- to long-term efficacy of DAI for a period of 4 weeks and found that DAI had possible short-term effects. Their conclusion is that long-term effects might not be expected, which seems in line with the present findings. In a study by Nordgren and Engstrom [6], DAIs were performed in the same manner as in the present study. Short-term effects on QoL were observed, but no long-term effects were present. However, the present analyses identified to a higher degree for women and participants with Alzheimer’s disease than for men and younger participants that median scores increased for separate items at post-test 2 compared to post-test 1. These findings indicate the possibility that certain groups can have more positive long-term effects than other groups. A recent review by Hu, Zhang [7] found that animal-assisted interventions can have short-term effects and possibly also long-term effects on psychological symptoms in people with cognitive impairment. Since the results in the present study and in other studies are inconclusive, more research about effects in different sub-groups is necessary. However, it is meaningful for the unique individual with dementia to have even short moments of wellbeing even though the results may not remain for a longer period.

Strengths and limitations

One strength of the present study was that the same dog handler completed all the interventions, and both the dog handler and the dogs are trained for their assignment and have extensive experiences with DAI in relation to people with dementia. Another strength was the relatively large number of participants. In addition, the instrument, QUALID, has been validated and is easy for nursing staff to complete.

There were some limitations in this study. The level of cognitive function was not measured, which means it is not possible to draw any firm conclusions about the effects of DAI in relation to cognitive function. Regardless, all participants had moderate to severe dementia; otherwise, they would not have been residents at the nursing homes in question. For practical reasons, a one group pretest – post-test study design was chosen to investigate the results of a treatment routine in the current municipality’s nursing homes. Consequently, there was no control group, and the researchers could not control for confounding factors in the residents’ environments. However, according to the pretest – post-test design, the residents served as their own controls. Nonetheless, it is possible that the participants’ QoL were affected by other factors such as other diseases, medications, changes at the nursing home such as new staff or new residents.

Another limitation is that the assessments of the residents’ QoL was made by proxy informants, i.e. by nursing staff, and the researchers had no influence over their assessments or values. There is also a possibility that the nursing staff’s assessments were influenced by their personal beliefs about DAI or the resident in question. An alternative to proxy-informed questionnaires is to use self-reports. Obviously, this could be complicated for people with severe dementia. Other measures, such as assessments of nutritional status, body weight (BMI), or observations of sleep patterns could conceivably complement the proxy ratings. Such measures, with the benefit of hindsight, could possibly have strengthened our results.

Conclusions

This study contributes to a growing scientific knowledge base about the use of therapy dogs in dementia care. The results indicated positive effects of DAIs on quality of life in nursing home residents with moderate to severe dementia. Most people would possibly agree that interactions between people and animals are effective and rewarding. Medical treatments are often used in dementia care to reduce symptoms, but they can also have negative side effects. Consequently, there is a need for nonpharmacological methods that can be used as alternatives or complements. However, fundamental questions remain. Is it worth the additional financial costs for training dogs and dog handlers? Moreover, is it worth putting both vulnerable and ill people and animals at risk of being injured or becoming sick? To answer these questions and others, more research is necessary.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This study was supported by grants from Demensförbundet and from Centre for Clinical Research Sörmland/Uppsala University.