Sim et al.

BMC Complementary Medicine and Therapies (2025) 25:303 BMC Complementary

[Link]
Medicine and Therapies

S YS T E M AT I C R E V I E W Open Access

Canine-assisted therapy in reducing stress

and anxiety levels of university students:
systematic review and meta-analysis
of randomized controlled trials
Shu Qing Sim1†, Zhiwei Liu1†, Zishuo Wu2, Sian Wang3 and Claudia Nisa1*

Abstract
Background Due to the high prevalence of mental health issues among university students worldwide, canine-
assisted therapy (CAT) has emerged as a potential intervention to reduce student stress and anxiety. This study
systematically reviews and meta-analyzes the effects of CAT on reducing stress and anxiety levels among university
students.
Methods Following the PRISMA 2020 guidelines, we conducted a systematic search across multiple databases (APA
PsycINFO, PubMed, Duke Libraries, CNKI, Wanfang, and Google Scholar) for randomized controlled trials published
in English and Chinese. Only studies incorporating professionally trained dogs and handlers were included. Two
reviewers (SS and ZL) independently extracted data, and the risk of bias was assessed using the Cochrane Risk of Bias
2 tool. Effect sizes (Hedges’ g) were pooled using a random-effects meta-analysis to account for the anticipated clinical
and methodological heterogeneity. Subgroup analyses were conducted to explore moderators such as intervention
duration, baseline stress levels, and control condition types. Publication bias was evaluated using funnel plots, Egger’s
test, and trim‑and‑fill analysis.
Results Of 290 identified studies, 15 met the inclusion criteria and 14 were included in the meta-analysis. The meta-
analysis yielded a statistically significant overall effect size of g = -0.67 (p <.001), indicating a moderate reduction
in stress and anxiety among university students receiving CAT. Forest plots revealed effect sizes ranging from
approximately − 1.34 to -0.13 across studies. Although substantial heterogeneity was observed, subgroup analyses
showed that CAT was significantly more effective for students with high baseline stress and anxiety levels. Funnel-plot
asymmetry suggested possible bias, but Egger’s test was non‑significant. Trim‑and‑fill analysis imputed two missing
studies, adjusting the pooled effect to g = -0.59, indicating the findings remain robust.
Conclusion CAT demonstrates promise in alleviating stress and anxiety among university students. Despite
methodological variations and potential publication bias, the findings suggest that CAT may offer a feasible and

†
Shu Qing Sim and Zhiwei Liu contributed equally to this work.
*Correspondence:
Claudia Nisa
cfn6@[Link]
Full list of author information is available at the end of the article

© The Author(s) 2025. Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0
International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you
give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the
licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or
exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit ​h​t​t​p​​:​/​/​​c​r​e​a​​t​i​​
v​e​c​​o​m​m​​o​n​s​.​​o​r​​g​/​l​​i​c​e​​n​s​e​s​​/​b​​y​-​n​c​-​n​d​/​4​.​0​/.
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 2 of 15

accessible approach to enhancing mental well-being within university environments. Further studies are necessary to
examine factors influencing methodological diversity and refine the integration of CAT within university settings.
Keywords Systematic review, Meta-Analysis, Canine-assisted therapy, University students, Stress level, Anxiety level,
Randomized controlled trials

Background [14, 15]. Therapy dogs and their handlers undergo spe-
Discussions are increasingly pointing to a growing cri- cialized training to ensure that interactions are benefi-
sis in student mental health [1]. In 2021 alone, approxi- cial and secure [15]. Often referred to as humans’ best
mately 73% of students in the United States reported friends, therapy dogs have been successfully incorporated
persistent psychological distress, a significant increase into various settings, including nursing homes, hospitals,
since 2013, reflecting a 50% rise in mental health chal- and schools, tailored to diverse populations [16]. Even
lenges within this population [2]. The causes of these briefly interacting, such as hugging, petting, and sitting
issues include academic workload pressure, loneliness, next to therapy dogs, can lower levels of stress hormones
inadequate sleep, and being away from family [3]. Pro- like epinephrine and norepinephrine, increase endor-
longed exposure to stress can contribute to the develop- phin levels and oxytocin, leading to decreased stress,
ment of more severe mental health conditions, further anxiety, and pain, improve mood, and better social bond-
increasing anxiety rates among university students [4]. ing [15]. Recent studies have shown that canine-assisted
Understanding different approaches to intervening in therapy effectively reduces stress among nurses [17] and
anxiety and stress is crucial, as these issues significantly enhances the quality of life for Alzheimer’s patients [18].
affect students’ daily functioning, life satisfaction, and Practically, canine-assisted therapy is relatively inex-
overall quality of life [5, 6]. When left unaddressed, stress pensive, as dogs are easily transportable and can engage
and anxiety are associated with risk factors such as aca- with a larger number of people during a single drop-in
demic burnout, relationship difficulties, decreased pro- session, making it a cost-effective option for university-
ductivity, and increased dropout rates [7]. In extreme based interventions [19]. A report indicated that 62%
cases, students may adopt maladaptive coping strategies, of surveyed universities in the United States provide
including excessive alcohol consumption, social with- animal-assisted therapy (AAT) programs, the majority
drawal, avoidance behaviors and even suicidal tendencies of which exclusively involved dogs [20]. This suggests
[5]. that canine-assisted therapy is a viable intervention for
In response to the impact of stress on both academic addressing the escalating mental health concerns in uni-
performance and overall well-being, universities have versities, particularly given the high burnout rate among
introduced various stress-reduction interventions such counselors and the financial implications of hiring more
as mindfulness programs, peer support groups, and on- professional counselors [12].
campus counseling services [8, 9]. However, the effective- Theoretically, canine-assisted therapy (CAT) is based
ness of these interventions remains inconsistent due to on social support theory, offering a source of social con-
underutilization and low student uptake. Peer support nectedness and combating feelings of loneliness and
may deter students from disclosing personal issues to isolation [21]. Social support, as defined by Thoits [22],
their peers, while mindfulness programs face operational refers to the social resources individuals can rely on
challenges: universities must invest in staff training, pro- when dealing with life challenges and stressors. Accord-
gram delivery, and materials, and students must dedi- ing to this theory, social support has several dimensions:
cate considerable time and effort to learn and maintain First, social support can be conceptualized as perceived
mindfulness practices [10, 11]. On-campus counseling support, reflecting an individual’s belief that support is
services are often strained under high demand, leading to available when needed [22]. Second, support can take
long waiting times for students and increased counselor various forms: instrumental, informational, or emo-
burnout and turnover [12]. Clinical interventions require tional. CAT primarily provides emotional support, which
qualified professionals and depend heavily on student involves expressions of sympathy, caring, esteem, value,
motivation and engagement, which can further limit or encouragement [23]. Additionally, such social relation-
their accessibility and impact [13]. Given these consid- ships extend beyond human interactions to encompass
erations, there is a need for interventions that are more connections with animals. As McNicholas and Collis
accessible and convenient for students, making canine- [24] noted, pets can complement existing human support
assisted therapy a promising option. networks and, in some cases, even serve as substitutes
Canine-assisted therapy is a complementary interven- for lacking human connections. Pets provide support
tion using professionally trained dogs to enhance individ- that releases individuals from relationship obligations,
uals’ physical, social, emotional, or cognitive functioning facilitates reorganization, and helps rebuild routines. In
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 3 of 15

this context, CAT presents a unique type of social sup- picture of how CAT may help support mental health in
port that nurtures emotional connection in individuals. this population.
Furthermore, having a supportive social network has a Therefore, this paper aims to systematically review the
direct impact on wellbeing, especially when it is large and effectiveness of canine-assisted therapy (CAT) in reduc-
is perceived to offer support, which can improve quality ing stress and anxiety among university students. To
of life and reduce stress [25]. Consequently, the dogs and address this gap, we conducted a meta-analysis focusing
handlers involved in CAT play a crucial role in engaging specifically on randomized controlled trials (RCTs) to
individuals in social interactions and offering social sup- establish a robust cause-and-effect relationship between
port, thereby buffering them from stress and anxiety [21]. CAT and mental health outcomes. By synthesizing evi-
Recognizing that heightened stress and anxiety can dence from these trials, we aim to provide a clearer
decrease students’ quality of life and performance [26], understanding of CAT’s potential in supporting student
CAT has emerged as a promising intervention across mental wellness.
diverse populations. Previous research has linked CAT
and animal-assisted interventions to mental health Methods
improvements such as reduced anxiety and pain-related PROSPERO registration
symptoms in children [27, 28], as well as decreased PTSD Our systematic review and meta-analyses were devel-
symptoms in mixed-age samples [29, 30]. These find- oped following the PRISMA 2020 statement [36]. As of
ings support CAT’s broader potential as an intervention April 2025, our PROSPERO registration record has ID
for promoting psychological well-being. Despite this, number CRD42024518138.
research specifically examining CAT’s impact within uni-
versity settings remains limited. We identified nine rele- Study selection
vant meta-analyses on animal-assisted therapy (Table 1), We focused on studies involving undergraduate and
but only one focused on higher education [31]. While graduate students of all genders, including those who
Huber et al. [31] explored a range of mental, cognitive, prefer not to disclose their gender. Students clinically
and physiological outcomes, with stress and anxiety as diagnosed with anxiety disorders were also eligible for
only part of the focus, their review also included general inclusion, given previous research demonstrating the
animal-assisted interventions (AAIs) involving various effectiveness of canine-assisted therapy (CAT) in treating
animals beyond dogs. Although that work provides help- anxiety [37].
ful insights, our study takes a more focused approach to This meta-analysis exclusively examined CAT. While
examine professionally facilitated CAT and its effects on conducting an overview of all similar treatments assisted
stress and anxiety in university students. By doing so, we by animals, we observed a standard confusion in the con-
aim to build on existing literature and provide a clearer cepts and experimental designs of these interventions.

Table 1 Past Meta-Analyses

Author, Year Age Range Outcomes Measured Animals Involved Intervention Type Effect Size (CI)
Nimer & Lundahl, 2007 [32] 0–65+ Autism spectrum Dogs, cats, rabbits, horses, Animal-assisted therapy Cohen’s d
behaviors aquatics, others 0.72 (0.23, 1.22)
Virues-Ortega et al., 2012 40+ Anxiety Dogs, cats, birds, rabbits, Animal-assisted therapy Hedge’s g
[33] dolphins, others -0.29 (-0.51, -0.07)
Charry-Sánchez et al., 2018 3–16 Gross motor function in Horses, dogs, elephants, Animal-assisted therapy Cohen’s d
[34] cerebral palsy dolphins 1.61 (-2.00, 5.23)
Germain et al., 2018 [29] 8–62 Trauma-related symptoms Horses, dogs, others Animal-assisted Hedge’s g
psychotherapy 0.86 (0.53, 1.18)
Waite et al., 2018 [35] 2–88 Pain, anxiety, distress Dogs Animal-assisted intervention Hedge’s g
1.65 (0.46, 2.83)
Hediger et al., 2021 [30] 4–86 Post-traumatic stress Dogs, horses, dolphins, Animal-assisted intervention SMD
disorder symptoms sheep, farm animals, seals, -0.26 (-0.56, 0.04)
others
Feng et al., 2021 [27] 3–17 Anxiety Dogs Canine-assisted therapy SMD
-0.19 (-0.44, 0.06)
Zhang et al., 2021 [28] 3–18 Pain-related symptoms Dogs Animal-assisted therapy MD
-0.53(-0.77, -0.30)
Huber et al., 2024 [31] 18–22 Anxiety Dogs, cats, fish, others Animal-assisted intervention Hedge’s g
-0.57, (-1.45, 0.31)
Note: Only 1 meta-analysis was identified to be done in university settings [31]; 5 studies included other animals rather than just therapy dogs; 2 studies restricted
the analyses to randomized controlled trials only [31, 34]; SMD = Standardized Mean Difference; MD = Mean Difference
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 4 of 15

For instance, it was often unclear whether the animals intervention led by a psychologist that teaches mindful-
and caregivers involved were professionally trained, or ness by encouraging participants to pet a dog’s fur with
whether handlers were present during sessions. Many the explicit goal of reducing anxiety would qualify as ani-
animal-assisted interventions share overlapping features mal-assisted psychotherapy [44]. In contrast, CAT does
but also differ in important ways. However, inconsistent not require clinical diagnoses or licensed therapists and is
terminology and vague criteria make it hard to clearly typically used in broader wellness or prevention contexts,
distinguish between different types of animal-assisted such as providing stress relief for university students who
therapy. may not have a diagnosed mental illness.
To address this, we used strict inclusion criteria to Broader animal-assisted interventions (AAI) were
define CAT in our study. For an intervention to be consid- excluded for two main reasons. First, AAI is an umbrella
ered CAT in our study, it must have two key criteria: (1) term that includes a range of practices, such as animal-
the incorporation of professionally trained dogs certified assisted activities (AAA), animal-assisted education
by relevant organizations; and (2) the dog handler must (AAE), and animal-assisted therapy (AAT) [46]. AAT
be present during the treatment [38]. The standardization itself can involve a variety of animals, including dogs,
of incorporating professionally trained, certified dogs and cats, guinea pigs, birds, dolphins, and others [33]. CAT
requiring the presence of a handler during CAT is cru- is a specific subtype of AAT that focuses exclusively on
cial for the safety, therapeutic efficacy, and ethical treat- professionally trained dogs. Because this study centers on
ment of the therapy dog [39]. Certified dogs are trained dog-based interventions, only those falling under CAT
to remain calm, responsive, and adaptable in various were included. Second, AAA and AAE may involve staff
therapeutic settings, reducing risks (e.g., agitation or bit- from various backgrounds, including educators, activ-
ing) associated with patient inappropriate behaviors such ity coordinators, or volunteers [47], while CAT requires
as abrupt gestures, rough handling, and direct interac- trained dog handlers to ensure safety and structure.
tions [39, 40]. The handler plays a key role in supervising Therefore, AAA and AAE, which may not consistently
interactions, recognizing and mitigating signs of stress in involve trained animals and handlers, were excluded
the dog, and upholding ethical standards during sessions. from this study.
Research shows that therapy dogs can experience stress The main comparator in this study was the control
or burnout if not properly managed, which may reduce group of students who were not exposed to the canine-
the therapy’s effectiveness and negatively impact animal assisted therapy intervention. Control group designs
welfare [39–41]. Professional organizations like IAHAIO may vary across studies, such as waitlist control, only
and AAII emphasize the importance of handler involve- handler control, watching videos about dogs’ control,
ment to ensure safe, ethical, and effective CAT sessions. and completing unrelated tasks. Our study also focused
Without these elements, both the outcomes and integrity exclusively on randomized controlled trials to estab-
of the intervention could be compromised [39, 42]. For lish a causal relationship between the intervention and
studies included in our analysis, if the presence of a han- its effects on students’ stress and anxiety levels. We did
dler was not explicitly stated, we assumed a handler was not include cross-sectional or within-subjects pre-post
present since this is standard practice in CAT. With these designs, but only studies that consist of canine-assisted
established criteria, even though some papers might not therapy as the treatment group and other forms of con-
name the intervention particularly as canine-assisted trol groups.
therapy, we still included them in our review as long as
they involved trained therapy dogs and professional Outcomes
handlers. The primary outcome was the decrease in reported stress
In contrast to CAT, several related interventions, such and anxiety levels in the group exposed to canine-assisted
as animal visitation programs (AVPs), animal-assisted therapy. Studies were included if they reported either or
psychotherapy (AAPT or AAP), and broader animal- both constructs. Measurement of stress and anxiety lev-
assisted interventions (AAI), were excluded from this els was based on standardized and validated self-rating
study. AVPs may involve informal interactions with ani- scales; we did not consider observers’ ratings. According
mals that are not professionally trained, which does not to the American Psychological Association [48], anxi-
meet this study’s requirement of incorporating certified ety is “an emotion characterized by feelings of tension,
therapy dogs [43]. AAPT, on the other hand, is a goal- worried thoughts, and physical changes like increased
directed, clinical intervention conducted by professional blood pressure.” Since anxiety could be categorized into
psychotherapists, where the animal is integrated as a key state and trait anxiety, we focused on state anxiety when
part of the therapeutic process aimed at reducing symp- studies reported anxiety levels separately. State anxiety
toms [44]. Therefore, it primarily focuses on patients with is conceived as a situational increase in feelings of inse-
clinically diagnosed mental illnesses [45]. For example, an curity, while trait anxiety is a relatively stable personality
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 5 of 15

characteristic [49]. With most canine-assisted therapy use of trained therapy dogs and handlers), and outcome
being implemented only occasionally in universities, measurement. A checklist based on mutually agreed
examining its effect on state anxiety would be more bene- criteria was used during screening (see supplementary
ficial for future recommendations, such as implementing materials), and any disagreements were resolved through
it before stressful events like exams. Meanwhile, stress is discussion. We restricted our search to randomized
defined as a state of worry or mental tension caused by controlled trials published in English or Chinese. Only
a difficult situation [50]. Additionally, to explore poten- peer-reviewed, published studies were considered; gray
tial moderators of the effectiveness of canine-assisted literature was excluded. Our search included studies pub-
therapy in reducing stress and anxiety, we conducted lished up to April 2025, and any articles published after
subgroup analyses based on variations in intervention this were not considered for inclusion.
duration, control condition type, baseline stress/anxiety
levels, and risk-of-bias ratings. Data extraction
Data extraction was completed by both reviewers inde-
Sources and search strategy pendently using an Excel sheet. We collected data on
We conducted a systematic literature search on April 10, study designs, study participants, sample sizes, treatment
2025, across several primary databases: APA PsycInfo, conditions, control conditions, baseline measurements,
PubMed, Duke Libraries, CNKI, and Wanfang. To maxi- reported post-measurements, duration of interventions,
mize sensitivity, we did an all-field search in each data- and the risk of bias. Risk of bias was assessed using the
base rather than limiting our query to titles and abstracts. Cochrane Risk of Bias 2 (RoB 2) tool [51]. Each study was
In addition, we used Google Scholar as a supplementary evaluated for randomization process, deviations from the
source to identify any additional peer-reviewed publica- intended intervention, handling of missing outcome data,
tions and hand-searched the reference lists of relevant consistency of outcome measurement, and selection of
systematic reviews and included studies for potentially reported results, with each domain rated as “low risk,”
eligible publications. All references were imported into “some concerns,” or “high risk.”
Covidence for deduplication and subsequent screening. To compute the effect sizes of each study, we primar-
Our search strategy focused on three core concepts: ily extracted post-means and standard deviations from
the population (university students), the intervention treatment and control groups, following the recommen-
(canine-assisted therapy and related animal-assisted dation by Lipsey and Wilson [52] for practical meta-
interventions), and the outcomes (stress and anxiety). analysis. If means and standard deviations were not
For example, we combined terms such as “college,” “uni- reported, we extracted t-test values instead. For stud-
versity,” “undergraduate,” and “graduate” for the popula- ies reporting results from ordinary least squares (OLS)
tion; “canine-assisted therapy,” “dog-assisted therapy,” regression models with a dummy-coded treatment vari-
“animal-assisted therapy,” and similar variants for the able (e.g., treatment = 1, control = 0), we extracted the
intervention; and “stress,” “stress levels,” “anxiety,” and unstandardized regression coefficient, its standard error,
“anxiety levels” for the outcomes. This approach ensured the standard deviation of the dependent variable, and
that studies addressing all three concepts were retrieved. group sample sizes to compute standardized effect sizes.
Detailed search strings for each database are provided in These were included only when the regression compared
the supplementary materials. An example of the search treatment versus control within an RCT. If data such
string is as follows: as sample sizes, intervention duration, or baseline val-
(“canine-assisted therapy” OR “canine therapy” OR ues were missing, we contacted the authors to request
CAT OR “dog-assisted therapy” OR “dog therapy” OR the information. Missing data with no author response
“canine-assisted intervention” OR “animal-assisted ther- were imputed through mean substitution; in the case of
apy” OR “animal therapy” OR “animal-assisted inter- subgroup analyses, such studies were excluded. Any dis-
vention” OR AAI) AND (college OR undergraduate OR agreements during the extraction process were resolved
graduate OR “higher education” OR “university stu- through discussion, and when needed, we consulted the
dents”) AND (stress OR “stress levels” OR anxiety OR supervising author (CN).
“anxiety levels”).
Two reviewers (SS and ZL) independently screened Data analysis
titles and abstracts based on predefined inclusion criteria. We conducted a random-effects meta-analysis using
Eligible studies were those mentioning canine-assisted Stata/SE 18.0 as we anticipated both clinical and meth-
therapy, targeting university students, and reporting odological heterogeneity among canine-assisted therapy
measures of stress or anxiety. Articles that passed this studies [31]. Given the considerable variability in inter-
initial screening underwent full-text review to verify vention protocols, participant characteristics, and out-
study design, intervention implementation (including the come measures in this field, the random-effects model is
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 6 of 15

most appropriate because it assumes that each study esti- Results

mates a different true effect [53], thus providing a more Study selection
conservative and generalizable estimate of the overall After running our search strategy across the selected
intervention effect. databases, we retrieved a total of 290 records (Fig. 1).
Effect sizes (Cohen’s d) were computed using the Automatic deduplication in Covidence identified 34
extracted post-intervention means and standard devia- duplicates, and an additional 2 duplicates were removed
tions, or t-test/regression statistics when necessary. The manually, leaving 254 articles for screening. Following the
practical meta-analysis effect size calculator [54] was title and abstract screening, 224 records were excluded
used to calculate effect sizes and standard errors for for not meeting the eligibility criteria (e.g., incorrect pop-
the meta-analysis. Subgroup analyses were conducted ulation, intervention, or study design), leaving 30 articles
to examine the impact of intervention duration, differ- for full-text review. Of these, 15 were further excluded
ences based on risk-of-bias levels, and the effect of con- during full-text review for reasons such as not using a
trol condition type on intervention effectiveness. Control randomized controlled trial design, not measuring stress
conditions were classified as passive if they involved no or anxiety outcomes (e.g., measured general mood only),
structured engagement (e.g., waitlist, routine academic involving mixed-species interventions (e.g., both cats and
activities) and as active if they involved structured or dogs), or using an intervention format outside the scope
interactive components (e.g., interacting with handler- of our review (e.g., a therapy dog passively wandering in a
only sessions, watching animal-related videos). To assess lecture hall). Specific reasons for exclusion at the full-text
the robustness of our findings for the intervention dura- stage are provided in the supplementary materials. No
tion subgroup, we also conducted a sensitivity analysis additional eligible studies were identified from the Chi-
by excluding multi-session studies and re-running the nese databases.
analysis. In total, 15 individual studies [13, 21, 57–69] met al.l
For both meta- and subgroup analyses, we reported inclusion criteria. One study [61] met al.l inclusion crite-
Hedges’ g, 95% confidence intervals, and p-values. The ria at the abstract and full-text screening stages; however,
chi-square test of group differences was also reported for it was excluded from the meta-analysis due to insuffi-
subgroup analyses to determine if effect sizes differed sig- cient statistical reporting. Specifically, the study did not
nificantly across groups. We considered Hedges’ g values provide post-intervention means, standard deviations,
of 0.20, 0.50, and 0.80 as small, medium, and large overall t-values, or enough information (e.g., standard error of
intervention effect sizes [55], respectively, and a p-value regression coefficients) to compute an effect size, and
less than 0.05 as significant. Heterogeneity across stud- our attempts to extract usable data did not meet our pre-
ies was assessed using the I² statistic. We interpreted I² defined extraction requirements. The remaining 14 stud-
values of 0–30% as insignificant, 30–60% as moderate, ies were included in the meta-analysis.
60–75% as substantial, and 75–100% as considerable het-
erogeneity [56]. Small-study bias was assessed through Study characteristics
Egger’s test, while potential publication bias was evalu- Most of the included studies were conducted in West-
ated using trim-and-fill analysis and funnel plots. All ern countries– 3 in Canada, 9 in the USA, and 3 in the
tests and analyses were conducted using Stata/SE 18.0’s UK– with a total sample of 1868 university students.
built-in functions; no external packages were used. Individual sample sizes ranged from 39 to 357 partici-
In addition, to account for differences in how baseline pants (M = 124, SD = 87). In most studies, the number
stress and anxiety were measured across studies, we stan- of female participants exceeded that of males, with an
dardized the baseline scores using z-scores. First, we cal- overall male-to-female ratio of 1:3 (Male N = 424; Female
culated the pooled mean baseline score by weighting each N = 1322), and the average age of participants was 19.61
study’s baseline score according to its sample size. We years (SD = 2.37). Most studies did not screen for prior
then estimated the pooled standard deviation by combin- anxiety or mental illness diagnoses; only one reported
ing the variances from all studies while adjusting for the that 33% of participants had previously been diagnosed
total number of studies. Finally, we computed a z-score with an anxiety disorder [57]. All interventions involved
for each study, representing how much its baseline stress live, in-person interactions with a therapy dog accompa-
and anxiety level deviated from the overall pooled mean. nied by a trained handler. Intervention session durations
Based on these z-scores, we categorized studies into low varied considerably, from as short as 1 min to as long as
(z < 0) and high (z > 0) baseline stress/anxiety groups in 120 min (M = 32, SD = 31.6). All studies allowed students
the later section to conduct a subgroup analysis. to interact freely with therapy dogs; students could pet,
sit with, or play with the dogs based on their preferences.
Most studies implemented a single-session interven-
tion; however, two studies administered the intervention
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 7 of 15

Fig. 1 PRISMA flow chart. A total of 290 records were identified from databases, and 36 studies were removed before screening. Of 254 studies that were
screened for titles and abstracts, 30 studies were eligible for full-text review. 15 studies were included in the review and 14 in the meta-analysis after
finalization

over 6 and 7 weeks, respectively [60, 64]. Control con- was the most common (N = 3), while a larger number of
ditions also varied. Examples included waitlist controls studies used active controls (e.g., engagement with han-
[64], watching videos of dogs [63], and completing the dlers, interaction with other participants, or viewing
Family Life-Space Diagram [13]. The waitlist condition non-animal-related presentations; N = 8) compared to
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 8 of 15

passive controls (e.g., continuation of daily routines such Main effect

as studying; N = 7). Results from random-effects meta-analysis revealed
Regarding outcome measures, 6 studies (40%) assessed a statistically significant overall effect size of g = -0.672
anxiety using the State-Trait Anxiety Inventory (STAI), (95% CI [-0.892, -0.452], z = -5.99, p <.001), indicating a
5 studies (33%) evaluated stress via versions of the Per- moderate effect. However, four specific studies [13, 58,
ceived Stress Scale (PSS), and the remaining studies 60, 68] had no statistically significant effects. Among all
used alternative tools such as momentary stress ratings studies, Trammell [67] demonstrated the strongest effect
or diagram-based measures like the Family Life-Space size of -1.342 (95% CI [-1.997, -0.687]), while Ward-
Diagram. Although the majority of studies reported Griffin et al. [68] reported the smallest significant effect
reductions in stress and anxiety levels in the interven- size of -0.129 (95% CI [-0.379, 0.122]). Additionally, high
tion group, four of them did not yield statistically signifi- heterogeneity was observed across the included stud-
cant results [13, 58, 60, 68]. Additional details on study ies (I2 = 72.81, p <.001), suggesting substantial variability
designs and individual study results are provided in the between intervention effects (Fig. 2).
Summary Table of Papers (Table 2).

Table 2 Summary table of papers (N = 15)

Paper Sample Intervention Control Measured Scale Estimates Esti-
Size Duration Outcome of Cohen’s mates
d of Risk
of Bias
Anderson & Brown, N = 89 35–45 min Interaction with other participants Anxiety State-Trait Anxiety -0.79 High
2021 [57] in a separate room without dogs Inventory
present
Barker et al., 2016 N = 57 15 min Completion of the Family Life- Perceived Perceived Stress Scale -0.38 High
[13] Space Diagram (attention-control) Stress
Barker et al., 2017 N = 74 15 min Completion of the Family Life- Perceived Family Life Space -0.36 High
[58] Space Diagram (attention-control) Stress Diagram
Binfet, 2017 [59] N = 163 20 min Engagement in a typical academic Perceived Perceived Stress Scale -0.32 High
activity, such as studying course Stress
material
Binfet et al., 2021 N = 284 20 min Interaction with a canine handler Momentary Momentary Stress -0.99 Me-
[21] only Stress (One item) dium
Carr & Pendry, 2025 N = 145 7 sessions Waitlist control Anxiety Beck Anxiety Inventory -0.03 Me-
[60] 2 h/session dium
Crump & Derting, N = 61 30 min Placement in another room Perceived Perceived Stress Scale NA High
2015 [61] Stress
Grajfoner et al., 2017 N = 132 20 min Interaction with a canine handler Anxiety State-Trait Anxiety -1.16 Me-
[62] only Inventory dium
Haefelin et al., 2020 N = 165 1 min Watching a video of playful dogs Anxiety State-Trait Anxiety -0.99 High
[63] on a provided laptop Inventory
Kivlen et al., 2022 N = 104 6 sessions Waitlist control Anxiety Patient-Reported Out- -0.40 Me-
[64] 35 min/session comes Measurement dium
Information System
Manville et al., 2022 N = 60 10 min Watching a PowerPoint presenta- Anxiety State-Trait Anxiety -0.65 High
[65] tion with neutral, non-animal Inventory
images
Spruin et al., 2020 N = 94 30 min Standard treatment for student Anxiety State-Trait Anxiety -1.22 High
[66] stress and anxiety: a one-on-one Inventory
session with a student advisor
Trammell, 2019 [68] N = 44 Not mentioned Placement in a similar room, but Perceived Perceived Stress Scale -1.34 Me-
without dogs present Stress dium
Ward-Griffin et al., N = 357 90 min Waitlist control Perceived Perceived Stress Scale -0.13 High
2018 [69] Stress
Williams et al., 2018 N = 39 12 min Engagement in normal pre-exam Anxiety State-Trait Anxiety -0.96 High
[67] routine, such as independent Inventory
studying
Note. N = Analyzed Participants; All studies were conducted in Western countries (e.g., Canada, USA, and UK). All interventions involved students interacting with
both dogs and handlers concurrently, while the control design varied across studies. Estimates of Cohen’s d were computed using post-intervention means and SDs
of the intervention and control groups or t-test/regression statistics
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 9 of 15

Fig. 2 Forest plot. Effect sizes were computed in Hedges’ g; CI = Confidence Interval; I2 = Heterogeneity Value; Vertical solid line represents the null effect;
Vertical dash line represents the weighted mean difference

Furthermore, the funnel plot (Fig. 3) raises concerns [-0.892, -0.452]), while the adjusted effect size after impu-
regarding potential publication bias, as 5 out of 14 stud- tation was − 0.592 (95% CI [-0.816, -0.368]). Although
ies fall outside the 95% confidence interval. For example, this adjusted effect size is slightly smaller in magnitude, it
the largest effect size of -1.34 by Trammell [67] origi- still reflects a moderate and statistically significant effect
nated from the study with the second smallest sample of canine-assisted therapy in reducing stress and anxi-
size (N = 44), suggesting possible small-study effects. To ety among university students. The adjusted funnel plot
formally assess this, we performed an Egger test, which (Fig. 4) showed a more symmetrical distribution, provid-
yielded a non-significant result (z = -1.27, p =.21). Even ing some evidence that the trim-and-fill method helped
though we did not find significant results for small study account for funnel plot asymmetry.
bias, it is still important to note that Egger’s test has low However, as suggested by Sterne et al. [51], the trim-
power to detect this bias when there are few studies. and-fill procedure assumes that funnel plot asymmetry
Thus, we complemented it with a trim-and-fill analysis is solely due to publication bias and may not perform
for a more comprehensive assessment of potential pub- well when substantial heterogeneity exists. Given the
lication bias. high heterogeneity observed in our meta-analysis, these
results should be interpreted with caution. Other fac-
Trim and fill analysis tors, such as clinical diversity or methodological vari-
The trim and fill analysis imputed two potentially miss- ability across studies, may also contribute to the initial
ing studies on the right side of the funnel plot, suggesting asymmetry.
that smaller or non-significant studies may be underrep-
resented. The observed effect size was − 0.672 (95% CI
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 10 of 15

Fig. 3 Funnel Plot. Evidence of potential publication bias is suggested, with 5 out of 14 studies falling outside the 95% confidence interval

Fig. 4 Adjusted Funnel Plot. A more symmetrical distribution is presented, with 2 missing studies imputed on the right side of the plot
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 11 of 15

Fig. 5 Summary of the risk of bias. No studies were identified to have an overall low risk of bias

Risk of bias on average compared to high-risk studies, this difference

Across 5 risk of bias domains and 75 total outcomes, was not significant.
17 of them were rated as low-risk, 43 were evaluated to
have some concerns, and 12 were identified as high-risk. Baseline differences
Overall, no studies were assessed as low-risk; 5 out of the Another potential bias in our analysis is that the samples
14 studies had some concerns, while the remaining were across studies were not homogeneous in their baseline
classified as high-risk. Similar to previous meta-analy- stress and anxiety levels, which may have either overes-
ses conducted on animal-assisted interventions, such as timated or underestimated the effect of CAT. To account
Huber et al. [31], we found that most studies could not for this bias, we divided the studies’ baseline z-scores
implement blinding procedures since the presence of into two groups: a low baseline stress/anxiety group
therapy canines during the intervention made it obvi- (z < 0) and a high baseline stress/anxiety group (z > 0). We
ous to both researchers and possibly the participants then conducted a subgroup analysis to examine whether
that they were in the treatment condition. This, in turn, CAT’s effectiveness varied depending on students’ initial
might have likely influenced the self-reported outcomes, stress and anxiety levels.
thereby causing many studies to be rated as “some con- Our analysis revealed that the pooled effect size for
cerns” and high-risk. The overview of the risk of bias the low stress/anxiety group was − 0.494 (95% CI [-0.778,
results can be referred to in Fig. 5, and specific details can -0.209]), which is smaller in magnitude than the pooled
be found in the supplementary materials. effect size of -0.909 (95% CI [-1.225, -0.593]) for the high
stress/anxiety group. The chi-square test of group differ-
Subgroup analyses ences indicated a statistically significant difference (Qb
Risk of bias = 3.67, p <.05) between the two groups. Furthermore,
To address potential biases arising from variations in the heterogeneity was more pronounced in the low stress/
risk of bias among studies included in our meta-analysis, anxiety group (I² = 73.17, p <.001) compared to the high
we conducted a subgroup analysis comparing the effect stress/anxiety group (I² = 59.65, p <.05). Overall, our
sizes of studies categorized as “some concerns” and subgroup analysis of baseline differences suggests that
“high-risk”. The pooled effect size for studies with some variations in students’ pre-intervention stress and anxiety
concerns was − 0.780 (95% CI [-1.232, -0.328]), while levels likely contributed to the heterogeneity of the main
for high-risk studies, it was − 0.607 (95% CI [-0.859, outcome.
-0.456]). We further evaluated the differences in effect
sizes between these groups using a chi-square test of Passive and active controls
group differences, yielding a non-significant result (Qb = To examine whether the type of control condition influ-
0.43, p =.513). In general, our findings suggest that while enced the effectiveness of canine-assisted therapy (CAT),
studies with some concerns showed a larger effect size we conducted a subgroup meta-analysis comparing
studies using passive versus active control conditions.
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 12 of 15

Interventions compared to passive controls (e.g., wait- differences were observed across subgroups. This sup-
list or typical study routines) yielded a pooled effect size ports the robustness of our findings since the inclusion of
of -0.503 (95% CI [-0.815, -0.191]), while those com- multi-session studies did not substantially alter conclu-
pared to active controls (e.g., interaction with a handler sions about intervention duration.
or engagement in attention-matched tasks) showed a
stronger effect size of -0.834 (95% CI [-1.046, -0.622]). Discussion
Although the test of subgroup differences did not reach Overall, our meta-analysis demonstrates that canine-
statistical significance (Qb = 2.96, p =.085), the results assisted therapy (CAT) effectively reduces university stu-
suggest a potential trend toward greater CAT effective- dents’ stress and anxiety, with a moderate pooled effect
ness when compared with more actively engaging control size. However, the substantial heterogeneity suggests that
conditions. CAT’s effectiveness may vary depending on the study
context. Although Egger’s test was not significant, the
Intervention duration fact that one of the smallest trials produced the largest
To explore potential differences in intervention effec- effect size points to possible small-study effects and calls
tiveness based on duration, we categorized sessions into for cautious interpretation.
three groups: 0–14 min (short), 15–29 min (medium), Subgroup analyses help clarify some of this variability.
and ≥ 30 min (long). Currently, there are no standardized Specifically, CAT produced a significantly larger effect
thresholds for session length in canine-assisted interven- among students with high baseline stress and anxiety
tions. However, prior research suggests that both brief compared to those with lower baseline levels. This find-
(5 min) and longer (20 min) sessions can significantly ing suggests that CAT may function more effectively as a
reduce stress [59], and one study found no significant targeted preventive intervention, particularly for individ-
differences in outcomes between sessions lasting 2-, 5-, uals already experiencing heightened psychological dis-
and 10 min [70]. Given this variability and the absence of tress. Additionally, there were no significant differences
established standards, we selected our cutoffs to reflect in effect sizes between passive control groups (e.g., wait-
common ranges reported in the literature. This approach list) and active control groups (e.g., attention-matched
may help clarify how session length relates to interven- activities), suggesting that the observed benefits of CAT
tion effectiveness and improve replicability in future are not solely due to novelty or attention effects.
studies. Although the shortest intervention duration
In the short-duration group (0–14 min), the pooled (0–14 min) group showed the largest effect size, sub-
effect size was − 0.883 (95% CI [-1.170, -0.596]), with no group differences by session length were not statistically
observed heterogeneity (I² = 0.00, p =.564), suggesting a significant. This implies that, while intervention length
consistent and strong effect of brief sessions in reducing might influence outcomes, no specific duration emerged
stress and anxiety. This consistency is notable given that as definitively superior. Instead, the consistent positive
one of the interventions [63] was as short as 1 min. The effects across all durations indicate CAT’s adaptability
medium-duration group (15–29 min) also demonstrated and its potential to be tailored flexibly to various pro-
a reduction in stress and anxiety, with a pooled effect size gram structures.
of -0.651 (95% CI [-1.001, -0.300]). However, substantial While potential publication bias cannot be entirely
heterogeneity was observed (I² = 74.52, p <.01), indicat- ruled out, given the funnel plot asymmetry and the impu-
ing variability across studies. Similarly, the long-duration tation of two potentially missing studies via trim-and-fill
group (≥ 30 min) yielded an effect size of -0.497 (95% analysis, the adjusted effect size remained statistically
CI [-0.906, -0.088]) with high heterogeneity (I² = 79.66, significant. This strengthens the overall conclusion that
p <.001), suggesting that longer sessions may be effective CAT is a promising intervention for improving student
but vary in impact. mental health.
A chi-square test of group differences indicated that From a theoretical standpoint, our findings align with
the differences between duration categories were not sta- social support theory. That is, interactions with therapy
tistically significant (Qb = 2.31, p =.283). Therefore, while dogs and handlers may serve as a source of social sup-
all durations show beneficial effects, no specific ses- port, resulting in reduced students’ anxiety and stress
sion length appeared as significantly more effective than levels [22]. This unique form of social support provided
others. by therapy dogs complements human support networks,
Two of the included studies implemented multi-session as suggested by McNicholas and Collis [14, 24]. The non-
interventions lasting several weeks [60, 64]. To examine judgmental, unconditional support facilitates emotional
their influence, we also conducted a sensitivity analysis relief, helping individuals reduce their stress and anxi-
excluding these studies. The effect sizes and direction ety levels. In particular, CAT facilitates emotionally safe,
of results remained consistent, and no meaningful
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 13 of 15

low-stakes social interactions, which may serve as an referred to in this analysis as canine-assisted therapy
important buffer against stress and anxiety. (CAT). Currently, there is no standardized theoreti-
In summary, this study contributes to the growing cal framework for evaluating the effectiveness of AAT/
body of literature supporting the effectiveness of CAT CAT, leading to wide variations in intervention protocols,
in reducing stress and anxiety, while also offering new treatment goals, and outcome measures. For example, the
insights into potential moderating factors such as base- duration and frequency of therapy dog interactions var-
line distress and intervention duration. It advances the ied considerably, ranging from brief one-minute sessions
field through rigorous inclusion of RCTs and by focusing to multi-week programs. Differences were also observed
specifically on university students, an underrepresented in the design of control conditions and outcome assess-
population in previous meta-analyses. ments, further contributing to methodological incon-
sistencies. Therefore, addressing these challenges by
Limitations and future directions standardizing research procedures, such as establishing
Included evidence consensus definitions and treatment guidelines, would
Similar to the previous AAI meta-analysis by Huber et help improve replicability and intervention reliability in
al. [31], a key limitation of the included studies is the future CAT studies.
predominance of female participants, with proportions It is also worth noting that although this study aligns
ranging from 60 to 95%. This imbalance in participant with social support theory, we did not directly measure
demographics may limit the generalizability of the find- social support factors such as emotional support, feel-
ings to the broader university student population, partic- ings of connectedness, or social isolation. Future studies
ularly given gender-based differences in stress perception should include explicit assessments of these constructs to
and coping styles [50, 71]. The overrepresentation of better understand the mechanisms by which CAT influ-
female participants is likely due to recruitment practices ences mental health. By examining the role of therapy
in fields such as psychology and nursing, which tend dog interactions in fostering a sense of connection and
to have higher proportions of female students. Future emotional relief, future studies can have deeper insight
research should aim to address this limitation by recruit- into the broader psychological benefits of CAT and clar-
ing more gender-diverse samples and by examining ify how social support mediates its effects on stress and
potential gender differences in CAT’s impact on stress anxiety reduction.
and anxiety reduction among university students.
Another limitation is the high risk of bias present in Implications
many studies, largely due to factors inherent to CAT The implications of these findings are important for uni-
interventions. First, carers and individuals delivering the versity administrators considering the implementation
interventions were often aware of participants’ group of CAT programs on their campuses. Despite the meth-
assignments, as therapy dog interactions are conspicu- odological limitations identified, CAT is still a promis-
ous (participants would see and interact with dogs), mak- ing solution for addressing the growing mental health
ing blinding infeasible. Second, outcome assessments challenges faced by university students, especially in the
may have been influenced by prior knowledge of CAT’s aftermath of the COVID-19 pandemic. Its cost-efficiency,
potential benefits, as many participants were psychology scalability, and high-popularity, combined with demon-
students who may have been familiar with the effects of strated effectiveness in reducing stress and anxiety, make
human-animal interactions. Third, a substantial num- CAT a feasible option for institutions aiming to promote
ber of studies lacked pre-registration, meaning analyses student well-being.
were not always conducted according to a pre-specified Universities may also consider integrating CAT into
plan finalized before unblinded outcome data were avail- broader, multifaceted mental health strategies by pairing
able, raising concerns about outcome reporting flexibil- it with existing services such as counseling, mindfulness
ity. Future research should adopt stronger trial designs, programs, or peer-led support groups. This integrated
including pre-registration of outcomes and the use of approach could encourage greater student engagement
validated, objective stress and anxiety biomarkers to and enhance the overall effectiveness of mental health
enhance internal validity. initiatives. Moreover, institutional investment in struc-
tured CAT programs (clear protocols for screening,
Review process supervision, and feedback) would contribute to the sus-
In addition, the studies included in this review exhib- tainability and standardization of these interventions,
ited high methodological heterogeneity, which may have ensuring long-term benefits for student populations.
influenced the interpretation of our results. We propose
that this heterogeneity is partly attributable to the broad
range of criteria for animal-assisted therapy (AAT),
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 14 of 15

Declarations
Conclusion
This meta-analysis provides strong evidence that canine- Ethics approval and informed consent
assisted therapy (CAT) is an effective intervention for Not applicable to secondary data.
reducing stress and anxiety among university students, Consent for publication
with an overall moderate effect size and consistent ben- Not applicable.
efits across session durations and control conditions.
Pre-Registration
These findings suggest that CAT can be successfully PROSPERO registration record: CRD42024518138.
implemented in higher education settings, especially for
students experiencing high levels of pre-existing stress or Competing interests
The authors declare no competing interests.
anxiety. Importantly, this study contributes to existing lit-
erature by isolating the effects of CAT within university Author details
1
student populations. While methodological limitations 2
Division of Social Sciences, Duke Kunshan University, Kunshan, China
Duke University, Duke Global Health Institute, Durham, USA
remain, the results support the inclusion of CAT as part 3
Lynch School of Education and Human Development, Boston College,
of a complementary suite of mental health interventions Boston, USA
within universities.
Future research should focus on improving study Received: 12 June 2024 / Accepted: 30 May 2025
design rigor, exploring long-term outcomes, and examin-
ing CAT’s differential effectiveness across diverse student
subgroups. Establishing standardized intervention pro-
References
tocols and directly assessing mechanisms, such as social 1. Bantjes J, Hunt X, Stein DJ. Anxious, depressed, and suicidal: crisis narratives
support and emotional bonding, will also be essential in university student mental health and the need for a balanced approach to
to optimizing CAT’s effectiveness. Ultimately, a deeper student wellness. Int J Environ Res Public Health. 2023;20(6):4859.
2. Fostering College Student Mental Health and Resilience. 2023; Available
understanding of CAT’s therapeutic potential will help from: ​h​t​t​p​​s​:​/​​/​w​w​w​​.​p​​s​y​c​​h​i​a​​t​r​y​.​​o​r​​g​/​n​​e​w​s​​-​r​o​o​​m​/​​a​p​a​​-​b​l​​o​g​s​/​​f​o​​s​t​e​​r​i​n​​g​-​c​o​​l​l​​e​g​e​​-​s​t​​
position this unique intervention as a valuable, accessi- u​d​e​n​​t​-​​m​e​n​t​a​l​-​h​e​a​l​t​h​-​a​n​d​-​r​e​s​i​l​i
ble, and student-centered approach to mental health care 3. Limone P, Toto GA. Factors that predispose undergraduates to mental issues:
A cumulative literature review for future research perspectives. Front Public
in higher education. Health, 2022. Volume 10–2022.
4. Tan GXD, et al. Prevalence of anxiety in college and university students: an
Abbreviations
umbrella review. J Affect Disorders Rep. 2023;14:100658.
AAI Animal-Assisted Interventions
5. Ribeiro ÍJS, et al. Stress and quality of life among university students: A
AAA Animal-Assisted Activities
systematic literature review. Health Professions Educ. 2018;4(2):70–7.
AAE Animal-Assisted Education
6. Tsitsas G, Nanopoulos P, Paschali AA. Life satisfaction, and anxiety levels
AAPT Animal-Assisted Psychotherapy
among university students. Creative Education; 2019.
AAT Animal-Assisted Therapy
7. Mofatteh M. Risk factors associated with stress, anxiety, and depres-
AVP Animal Visitation Program
sion among university undergraduate students. AIMS Public Health.
CAT Canine-Assisted Therapy
2021;8(1):36–65.
RCT Randomized Controlled Trial
8. J., D. Colleges Are Getting Creative to Support Mental Health. 2023; Available
from: ​h​t​t​p​​s​:​/​​/​t​i​m​​e​.​​c​o​m​​/​6​3​​1​6​2​1​​7​/​​c​o​l​​l​e​g​​e​-​m​e​​n​t​​a​l​-​h​e​a​l​t​h​-​p​r​o​g​r​a​m​s​/
9. Frazier P, et al. Understanding stress as an impediment to academic perfor-
Supplementary Information mance. J Am Coll Health. 2019;67(6):562–70.
The online version contains supplementary material available at ​h​t​t​p​​s​:​/​​/​d​o​i​​.​o​​r​ 10. Mikolajczak-Degrauwe K, et al. Strengths, weaknesses, opportunities and
g​/​​1​0​.​​1​1​8​6​​/​s​​1​2​9​0​6​-​0​2​5​-​0​4​9​5​5​-​2. threats of peer support among disadvantaged groups: A rapid scoping
review. Int J Nurs Sci. 2023;10(4):587–601.
Supplementary Material 1 11. Wagner AP, et al. Cost-effectiveness of providing university students
with a mindfulness-based intervention to reduce psychological distress:
economic evaluation of a pragmatic randomised controlled trial. BMJ Open.
Acknowledgements 2023;13(11):e071724.
Not applicable. 12. ME F. The Mental Health Crisis on College Campuses 2023; Available from: ​h​t​t​
p​​s​:​/​​/​w​w​w​​.​n​​e​a​.​​o​r​g​​/​n​e​a​​-​t​​o​d​a​​y​/​a​​l​l​-​n​​e​w​​s​-​a​​r​t​i​​c​l​e​s​​/​m​​e​n​t​​a​l​-​​h​e​a​l​​t​h​​-​c​r​​i​s​i​​s​-​c​o​​l​l​​e​g​e​-​c​
Author contributions a​m​p​u​s​e​s
SS analyzed and interpreted the data from the meta-analysis and subgroup 13. Barker SB, et al. A randomized Cross-over exploratory study of the effect of
analysis results. ZL was a major contributor to drafting the manuscript. Both visiting therapy dogs on college student stress before final exams. Anthro-
SS and ZL contributed equally to study selection, screening, evaluating the zoös. 2016;29(1):35–46.
levels of bias, and the overall design of the study. The second authors, ZW and 14. McDowall S, et al. Understanding the role of therapy dogs in human health
SW, reviewed and edited the statistical results and manuscript. CN supervised promotion. Int J Environ Res Public Health. 2023;20(10):5801.
the scientific quality of the entire meta-analytic process and revised the final 15. Marcus DA. The science behind Animal-Assisted therapy. Curr Pain Headache
manuscript. Rep, 2013. 17(4).
16. K C. Learning about the experiences of canine-assisted therapy handlers
Funding 2021; Available from: ​h​t​t​p​​s​:​/​​/​c​t​a​​r​c​​h​i​v​​e​.​c​​o​u​n​s​​e​l​​i​n​g​​.​o​r​​g​/​2​0​​2​1​​/​0​8​​/​l​e​​a​r​n​i​​n​g​​-​a​b​​o​
The current study is not being funded. u​t​​-​t​h​e​​-​e​​x​p​e​​r​i​e​​n​c​e​s​​-​o​​f​-​c​​a​n​i​​n​e​-​a​​s​s​​i​s​t​e​d​-​t​h​e​r​a​p​y​-​h​a​n​d​l​e​r​s​/
17. Machová K, et al. Canine-Assisted therapy improves Well-Being in nurses. Int J
Data availability Environ Res Public Health. 2019;16(19):3670.
No datasets were generated or analysed during the current study.
Sim et al. BMC Complementary Medicine and Therapies (2025) 25:303 Page 15 of 15

18. Sánchez-Valdeón L, et al. Canine-Assisted therapy and quality of life in people 45. Jones MG, Rice SM, Cotton SM. Incorporating animal-assisted therapy in
with Alzheimer-Type dementia: pilot study. Frontiers in Psychology; 2019. p. mental health treatments for adolescents: A systematic review of canine
10. assisted psychotherapy. PLoS ONE. 2019;14(1):e0210761.
19. Narvekar HN, Narvekar HN. Canine-assisted therapy in neurodevelopmental 46. Animal Assisted. Intervention. 2017; Available from: ​h​t​t​p​​s​:​/​​/​a​a​i​​-​i​​n​t​.​​o​r​g​​/​a​a​i​​/​a​​n​i​
disorders: A scoping review. Eur J Integr Med. 2022;50:102112. m​​a​l​-​​a​s​s​i​​s​t​​e​d​-​i​n​t​e​r​v​e​n​t​i​o​n​/
20. Haggerty JM, Mueller MK. Animal-assisted stress reduction programs in 47. Bert F, et al. Animal assisted intervention: A systematic review of benefits and
higher education. Innov High Educ. 2017;42(5–6):379–89. risks. Eur J Integr Med. 2016;8(5):695–706.
21. Binfet J-T, Green FL, Draper ZA. The importance of client–canine contact in 48. American Psychological Association. Anxiety. 2022; Available from: ​h​t​t​p​​s​:​/​​/​w​
canine-assisted interventions: A randomized controlled trial. Anthrozoös. w​w​​.​a​​p​a​.​​o​r​g​​/​t​o​p​​i​c​​s​/​a​n​x​i​e​t​y
2021;35(1):1–22. 49. Meijer J. Stress in the relation between trait and state anxiety. Psychol Rep.
22. Thoits PA. Stress, coping, and social support processes: where are we?? 2001;88(3suppl):947–64.
we?hat next?? J Health Soc Behav. 1995;35:53. 50. Merz CJ, Wolf OT. Examination of cortisol and state anxiety at an academic
23. Kort-Butle L. Social support theory: the encyclopaedia of juvenile delin- setting with and without oral presentation. Stress. 2015;18(1):138–42.
quency and justice. Wiley Blackwell; 2017. 51. Sterne J, Egger; M, Moher D. Chap. 10: addressing reporting biases. Cochrane
24. McNicholas J. G.M. Collis 2006 Animals as social supports: insights for Under- handbook for systematic reviews of intervention. Version 5.0. 1. The Cochrane
standing animal-assisted therapy. Handb animal-assisted Therapy: Theoretical Collaboration; 2008.
Found Guidelines Pract 2 49–72. 52. Lipsey MW, Wilson DB. Practical meta-analysis. Practical meta-analysis. Thou-
25. Panayiotou G, Karekla M. Perceived social support helps, but does not buffer sand Oaks, CA, US: Sage Publications, Inc. ix; 2001. pp. 247–ix.
the negative impact of anxiety disorders on quality of life and perceived 53. Dettori JR, Norvell DC, Chapman JR. Fixed-Effect vs Random-Effects models
stress. Soc Psychiatry Psychiatr Epidemiol. 2013;48(2):283–94. for Meta-Analysis: 3 points to consider. Global Spine J. 2022;12(7):1624–6.
26. Córdova P, Hernán PGG, Naranjo Mejía, La I, Chacón AG. Alberto Sanjinés 54. Wilson DB. Practical meta-analysis effect size calculator (Version 2023.11.27).
Unzueta. Academic stress as a predictor of mental health in university students. 2023.
2023; Available from: ​h​t​t​p​​s​:​/​​/​w​w​w​​.​t​​a​n​d​​f​o​n​​l​i​n​e​​.​c​​o​m​/​​d​o​i​​/​f​u​l​​l​/​​h ​ ​t​t​​p​s​:​/​​/​d​​o​i​.​​o​r​g​​/​1​ 55. Brydges CR. Effect size guidelines, sample size calculations, and statistical
0​.​​1​0​​8​0​/​​2​3​3​​1​1​8​6​​X​.​​2​0​2​3​.​2​2​3​2​6​8​6 power in gerontology. Innov Aging. 2019;3(4):igz036.
27. Feng Y, et al. Effects of Animal-Assisted therapy on hospitalized children 56. Deeks JJ et al. Analysing data and undertaking meta-analyses, in Cochrane
and teenagers: A systematic review and Meta-Analysis. J Pediatr Nurs. Handbook for Systematic Reviews of Interventions. 2019. pp. 241–284.
2021;60:11–23. 57. Anderson D, Brown S. The effect of animal-assisted therapy on nursing stu-
28. Zhang Y, et al. Effectiveness of animal-assisted therapy on pain in children: A dent anxiety: A randomized control study. Nurse Educ Pract. 2021;52:103042.
systematic review and meta-analysis. Int J Nurs Sci. 2021;8(1):30–7. 58. Barker SB, et al. The effect of a Canine-assisted activity on college stu-
29. Germain SM, et al. Animal-assisted psychotherapy and trauma: A Meta-analy- dent perceptions of family supports and current stressors. Anthrozoös.
sis. Anthrozoös. 2018;31(2):141–64. 2017;30(4):595–606.
30. Hediger K, et al. Effectiveness of animal-assisted interventions for children 59. Binfet J-T. The effects of Group-Administered canine therapy on uni-
and adults with post-traumatic stress disorder symptoms: a systematic review versity students’ wellbeing: A randomized controlled trial. Anthrozoös.
and meta-analysis. Eur J Psychotraumatology. 2021;12(1):1879713. 2017;30(3):397–414.
31. Huber A, et al. Animal-Assisted interventions improve mental, but not 60. Carr AM. Pendry Effects of an Animal-Assisted Drop-In program on First-Year
cognitive or physiological health outcomes of higher education stu- university students’ trajectory of psychological wellbeing. Pets. 2025;2. ​h​t​t​p​s​:​​​/​​/​d​o​​
dents: a systematic review and Meta-analysis. Int J Mental Health Addict. i​.​​o​r​​g​​/​​1​0​​.​3​3​​​9​0​​/​p​e​t​s​2​0​1​0​0​0​8.
2024;22(3):1597–628. 61. Crump C, Derting TL. Effects of pet therapy on the psychological and physi-
32. Nimer J, Lundahl B. Animal-Assisted therapy: A Meta-Analysis. Anthrozoös. ological stress levels of first-year female undergraduates. North Am J Psychol.
2007;20(3):225–38. 2015;17(3):575–90.
33. Virués-Ortega J, et al. Effect of animal-assisted therapy on the psychological 62. Grajfoner D, et al. The effect of dog-assisted intervention on student well-
and functional status of elderly populations and patients with psychiatric being, mood, and anxiety. Int J Environ Res Public Health. 2017;14(5):483.
disorders: a meta-analysis. Health Psychol Rev. 2012;6(2):197–221. 63. Haefelin N, et al. Anxiety reduction in college students after brief inter-
34. Charry-Sánchez JD, Pradilla I, Talero-Gutiérrez C. Effectiveness of Animal- action with a therapy dog or animatronic dog. North Am J Psychol.
Assisted therapy in the pediatric population: systematic review and Meta- 2020;22(3):411–411.
Analysis of controlled studies. J Dev Behav Pediatr. 2018;39(7):580–90. 64. Kivlen C, et al. Canine-assisted intervention effects on the well-being of
35. Waite TC, Hamilton L, O’Brien W. A meta-analysis of animal assisted interven- health science graduate students: A randomized controlled trial. Am J Occup
tions targeting pain, anxiety and distress in medical settings. Complement Therapy. 2022;76(6):7606205120.
Ther Clin Pract. 2018;33:49–55. 65. Manville K, Coulson M, Reynolds G. Canine-Assisted intervention reduces
36. Page MJ et al. The PRISMA 2020 statement: an updated guideline for report- anxiety and stress in higher education students: A randomized controlled
ing systematic reviews. BMJ, 2021: p. n71. trial. Volume 32. Society & Animals; 2022. pp. 156–77. 2.
37. Barker SB, Dawson KS. The effects of Animal-Assisted therapy on anxiety rat- 66. Spruin E, et al. The effects of a therapy dog vs mindfulness vs a stu-
ings of hospitalized psychiatric patients. Psychiatric Serv. 1998;49(6):797–801. dent advisor on student anxiety and well-being. J Furth High Educ.
38. Standards of Practice in. Animal-Assisted Interventions 2017; Available from: ​ 2021;45(5):588–600.
h​t​t​p​​s​:​/​​/​t​h​e​​r​a​​p​y​a​​n​i​m​​a​l​s​t​​a​n​​d​a​r​d​s​.​o​r​g​/ 67. Trammell JP. Therapy dogs improve student affect but not memory. Anthro-
39. Mignot A, et al. Handlers’ representations on therapy dogs’ welfare. Animals. zoös. 2019;32(5):691–9.
2022;12(5):580. 68. Ward-Griffin E, et al. Petting away pre‐exam stress: the effect of therapy dog
40. Glenk LM. Current perspectives on therapy dog welfare in animal-assisted sessions on student well‐being. Stress Health. 2018;34(3):468–73.
interventions. Animals. 2017;7(2):7. 69. Williams C, et al. An animal-assisted intervention’s influence on graduate
41. Haubenhofer DK, Kirchengast S. Physiological arousal for companion dogs students’ stress and anxiety prior to an examination. Open Access Libr J.
working with their owners in animal-assisted activities and animal-assisted 2018;5(9):1–16.
therapy. J Appl Anim Welfare Sci. 2006;9(2):165–72. 70. Manville K, Coulson M, Reynolds G. An exploratory randomised controlled trial
42. IAHAIO. The IAHAIO Definitions for Animal-Assisted Intervention and Guide- comparing the effectiveness of different duration of canine-assisted interventions
lines for Wellness of Animals Involved in AAI… 2019; Available from: ​h​t​t​p​​s​:​/​​/​o​ in higher education students. Human-Animal Interactions, 2023. 2023(2023).
t​a​​u​s​​.​c​o​​m​.​a​​u​/​p​u​​b​l​​i​c​a​​s​s​e​​t​s​/​8​​0​c​​f​d​5​​2​3​-​​2​0​3​0​​-​e​​a​1​1​​-​9​4​​0​3​-​0​​0​5​​0​5​6​​b​e​1​​3​b​5​/​​A​A​​I​I​-​​S​t​a​​ 71. Taylor MK, et al. Sex differences in cardiovascular and subjective stress
n​d​a​r​​d​s​​-​o​f​-​P​r​a​c​t​i​c​e​.​p​d​f reactions: prospective evidence in a realistic military setting. Stress.
43. Crossman MK, Kazdin AE. Animal visitation programs in colleges and universi- 2014;17(1):70–8.
ties. Elsevier; 2015. pp. 333–7.
44. Animal-Assisted Psychotherapy Theory, Issues, and Practice, ed. N. Parish-
Plass. 2013: Purdue University Press. Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.