Effects of Therapeutic Horseback-Riding Program on Social and Communication Skills in Children with Autism Spectrum Disorder: A Systematic Review and Meta-Analysis

Animal-assisted therapy has become a fast-growing and effective approach for remediating core impairments of children with ASD; however, recent systematic review studies on the effects of AAT in children with ASD have some limitations, including referral to a variety of animal-assisted interventions rather than to horseback-riding therapy alone and the absence of any meta-analysis in systematic reviews. A complete systematic review of the studies that describe the use of THR as an intervention is needed to specifically target the core impairments of children with ASD. The purpose of this study was to employ the systematic review method to synthesize research findings regarding the effects of THR programs on the social interaction and communication skills of children with ASD. We conducted a structured search in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We searched for potentially relevant studies in five databases (Web of Science, PubMed, CINAHL, Scopus, and SPORTDiscus) from inception until February 2022. In addition, we manually searched the bibliographies of the included studies to find articles that might otherwise have been missed. We considered articles eligible or ineligible if they satisfied specific inclusion or exclusion criteria. Our results showed that the THR program is an effective direct and alternative therapeutic program that can considerably improve the social behaviors and communication skills of children with ASD and can effectively impact autistic impairments in areas such as social awareness, social cognition, social motivation, and social communication. These findings are in line with those of previous studies; however, we did not find statistical evidence of any effect of THR on the autistic behaviors of irritability, stereotypy, and inappropriate speech. In conclusion, the findings produced by this meta-analysis study provide evidence that THR programs can considerably improve the social behaviors and communication skills of children with ASD.


Introduction
According to a 2018 CDC report by the Johns Hopkins School of Public Health, the number of individuals with autism spectrum disorder (ASD) has dramatically increased in recent years; it has nearly tripled, from 0.67% to 1.85% compared with the figure for 2000 [1,2] (Johns Hopkins report, 2020; CDC 2018 report). ASD typically impairs interpersonal social interaction and communication, and results in stereotypic behaviors. In children, these autistic behaviors seriously affect their social life with their families and

Inclusion and Exclusion Criteria
We considered articles to have met inclusion criteria if they involved: (i) diagnosed children with autism; (ii) randomized controlled studies (RCTs); (iii) horse-assisted therapy in the intervention group; (iv) a usual-care or wait-list condition for the control group; (v) measurements of social function; (vi) publication in the English language.
Exclusion criteria were: (i) the use of animals other than horses for trials; (ii) insufficient information for calculating the effect size (ES); (iii) case studies, observational studies, or review articles; (iv) duplicated study.

Data Extraction and Quality Assessment
We extracted detailed information concerning the author and publication year, characteristics of the study population such as sample size and ages of those involved, intervention protocols, diagnostic criteria, assessment tools, and outcomes.
The quality of each study was assessed by two independent authors using the Physiotherapy Evidence Database (PEDro) scale [40]. This assessment scale has 11 domains: eligibility criteria, random allocation, concealed allocation, similar measures between groups at baseline, instructor blinding, assessor blinding, participant blinding, more than 85% dropout rate, intention-to-treat analysis, statistical comparison between groups, and ≥1 key outcome estimated. Each item is scored as 0 (absent) or 1 (present). The total score ranges from 0 to 10 points. Study quality is classified as excellent (9-10 points), good (6-8 points), fair (4-5 points), or poor (<4 points) based on the assessment.

Statistical Analysis
We conducted a meta-analysis to measure the aggregative effect size (ES) of the effect of horse-assisted therapy on social function. In the analysis, we used standard mean differences (SMDs) to express the ESs by calculating the mean change from baseline to post intervention for the intervention and control groups. To analyze variables across studies, we used a random effects model with 95% confidence intervals (CIs) in the overall ESs calculated. We categorized the ESs as small (0.2-0.49), moderate (0.50-0.79), or large (≥0.8) based on the Higgins recommendation [41]. We used the I 2 test to assess study heterogeneity, and the three cut-off points for the levels of low, moderate, and high heterogeneity were I 2 = 25%, I 2 = 50%, and I 2 = 75%, respectively [42]. We conducted all statistical analyses using the Comprehensive Meta-Analysis program (version 2.2). We used a significance level of p < 0.05 in all analyses.

Search Results
The flowchart of selection studies is displayed in Figure 1. Initially, we identified 136 potentially suitable studies from the database and 2 studies from manual retrieval. Of these, we excluded 33 because of duplication, so we selected 105 full-text studies for a final review of titles and abstracts. We then reviewed 28 of these studies for eligibility by a reading of their full texts. Finally, we selected 5 studies [31,[43][44][45][46] for meta-analysis.
based on the Higgins recommendation [41]. We used the I 2 test to assess study he neity, and the three cut-off points for the levels of low, moderate, and high hetero were I 2 = 25%, I 2 = 50%, and I 2 = 75%, respectively [42]. We conducted all statistical a using the Comprehensive Meta-Analysis program (version 2.2). We used a sign level of p < 0.05 in all analyses.

Search Results
The flowchart of selection studies is displayed in Figure 1. Initially, we identi potentially suitable studies from the database and 2 studies from manual retrie these, we excluded 33 because of duplication, so we selected 105 full-text studi final review of titles and abstracts. We then reviewed 28 of these studies for eligib a reading of their full texts. Finally, we selected 5 studies [31,[43][44][45][46] for meta-analy  Table 1 shows detailed characteristics of the five included studies. Their years lication ranged from 2009 to 2019. A total of 240 children with autism were includ their ages ranged from 6 to 16 years old. All children in the intervention group offered horse-based therapy, which involved 45-70 min sessions of horse-assisted 1-2 times per week, for 7-24 weeks. Those in control groups were placed on a wai  Table 1 shows detailed characteristics of the five included studies. Their years of publication ranged from 2009 to 2019. A total of 240 children with autism were included, and their ages ranged from 6 to 16 years old. All children in the intervention groups were offered horse-based therapy, which involved 45-70 min sessions of horse-assisted riding, 1-2 times per week, for 7-24 weeks. Those in control groups were placed on a waiting list or participated in regular activities. The measured outcomes included assessments of social function (social communication, social cognition, social awareness, and social motivation) and aberrant behavior (irritability, lethargy/social withdrawal, stereotypy, hyperactivity, and inappropriate speech). Our appraisal of study quality is presented in Table 2. We identified four good-quality studies (RCTs) and one fair-quality study (n-RCT) among the included studies. All studies used concealed allocation, subject blinding, and therapist blinding. One study used assessor blinding. The remaining items (similarity in key measures at baseline, and comparison of more than one outcome) were described in each study. Studies were classified as excellent (9-10), good (6)(7)(8), fair (4)(5), or poor (<4).

Meta-Analysis Results of Social Function
We assessed the severity of symptoms associated with autism spectrum disorders (ASDs) using the SRS. The pooled meta-analysis results from the selected studies showed that horse-assisted therapy had a significant effect on the social functioning of children with ASD compared with the control group (Table 3). Among the five studies analyzed, the pooled meta-analysis results from the three studies in Figure 2  The pooled meta-analysis results from the four studies in Figure 3 showed that horseassisted therapy significantly improved social awareness in children with autism (SMD = −0.93, 95% CI [−1.62, −0.24], p < 0.001).  The pooled meta-analysis results from the four studies in Figure 3 showed that horse-assisted therapy significantly improved social awareness in children with autism (SMD = −0.93, 95% CI [−1.62, −0.24], p < 0.001).

Figure 2.
Forest plot showing the effects of horse-assisted therapy vs. control on social commun tion [43,44,46].
The pooled meta-analysis results from the four studies in Figure 3 showed that ho assisted therapy significantly improved social awareness in children with autism (SM −0.93, 95% CI [−1.62, −0.24], p < 0.001).     The pooled meta-analysis results from the four studies in Figure 3 showed that ho assisted therapy significantly improved social awareness in children with autism (SM −0.93, 95% CI [−1.62, −0.24], p < 0.001).

Meta-Analysis Results for Maladaptive Behaviors
The ABC-C scale was used to assess five maladaptive behavior problems: irritabi lethargy, hypersensitivity, stereotypy, and inappropriate speech.
Among the five studies included, the pooled meta-analysis results from the two st ies (in Figure 6) and three studies (in Figure 7) showed that horse-assisted therapy did

Meta-Analysis Results for Maladaptive Behaviors
The ABC-C scale was used to assess five maladaptive behavior problems: irritability, lethargy, hypersensitivity, stereotypy, and inappropriate speech.
Among the five studies included, the pooled meta-analysis results from the two studies (in Figure 6) and three studies (in Figure 7) showed that horse-assisted therapy did not significantly improve irritability (SMD = −0. 22

Meta-Analysis Results for Maladaptive Behaviors
The ABC-C scale was used to assess five maladaptive behavior problems: irritabi lethargy, hypersensitivity, stereotypy, and inappropriate speech.
Among the five studies included, the pooled meta-analysis results from the two s ies (in Figure 6) and three studies (in Figure 7) showed that horse-assisted therapy did significantly improve irritability (SMD = −0. 22   Among the five studies included, the pooled meta-analysis results from the t studies in Figure 8 showed that horse-assisted therapy produced a significantly pos effect on lethargy in children with ASD (SMD = −0. 32

Meta-Analysis Results for Maladaptive Behaviors
The ABC-C scale was used to assess five maladaptive behavior problems: irritab lethargy, hypersensitivity, stereotypy, and inappropriate speech.
Among the five studies included, the pooled meta-analysis results from the two s ies (in Figure 6) and three studies (in Figure 7) showed that horse-assisted therapy did significantly improve irritability (SMD = −0. 22   Among the five studies included, the pooled meta-analysis results from the t studies in Figure 8 showed that horse-assisted therapy produced a significantly pos effect on lethargy in children with ASD (SMD = −0.32, 95% CI [−0.64, −0.001], p = 0.04 Gabriels, 2015 -  Among the five studies included, the pooled meta-analysis results from the three studies in Figure 8  The pooled meta-analysis results from the three studies in Figure 9 showed horse-assisted therapy produced a significantly positive effect on hypersensitivity in c dren with ASD (SMD = −0.70, 95% CI [−1.29, −0.10], p = 0.02).  The pooled meta-analysis results from the three studies in Figure 9 showed that horseassisted therapy produced a significantly positive effect on hypersensitivity in children with ASD (SMD = −0.70, 95% CI [−1.29, −0.10], p = 0.02). The pooled meta-analysis results from the three studies in Figure 9 showed horse-assisted therapy produced a significantly positive effect on hypersensitivity in c dren with ASD (SMD = −0.70, 95% CI [−1.29, −0.10], p = 0.02). The pooled meta-analysis results from the three studies in Figure 10 showed horse-assisted therapy did not produce a significant improvement in inappropr speech behavior in children with ASD (SMD = −0.28, 95% CI [−0.59, 0.04], p = 0.08).

Discussion
Our purpose in this study was to use the techniques of a systematic review and m analysis to synthesize the findings of previous studies regarding the effects of THR p grams on the social interaction and communication skills of children with ASD. We lected 4 good-quality studies and 1 fair-quality study from a total of 138 initially identi AAI articles. A total of 240 participants were included and analyzed for this study. results of the meta-analysis provided evidence that THR programs can considerably prove the social interaction and communication skills of children with ASD. Although  The pooled meta-analysis results from the three studies in Figure 10 showed that horse-assisted therapy did not produce a significant improvement in inappropriate speech behavior in children with ASD (SMD = −0.28, 95% CI [−0.59, 0.04], p = 0.08). The pooled meta-analysis results from the three studies in Figure 9 showed t horse-assisted therapy produced a significantly positive effect on hypersensitivity in c dren with ASD (SMD = −0.70, 95% CI [−1.29, −0.10], p = 0.02). The pooled meta-analysis results from the three studies in Figure 10 showed t horse-assisted therapy did not produce a significant improvement in inappropr speech behavior in children with ASD (SMD = −0.28, 95% CI [−0.59, 0.04], p = 0.08).

Discussion
Our purpose in this study was to use the techniques of a systematic review and m analysis to synthesize the findings of previous studies regarding the effects of THR p grams on the social interaction and communication skills of children with ASD. We lected 4 good-quality studies and 1 fair-quality study from a total of 138 initially identif AAI articles. A total of 240 participants were included and analyzed for this study. T results of the meta-analysis provided evidence that THR programs can considerably prove the social interaction and communication skills of children with ASD. Although

Discussion
Our purpose in this study was to use the techniques of a systematic review and meta-analysis to synthesize the findings of previous studies regarding the effects of THR programs on the social interaction and communication skills of children with ASD. We selected 4 good-quality studies and 1 fair-quality study from a total of 138 initially identified AAI articles. A total of 240 participants were included and analyzed for this study. The results of the meta-analysis provided evidence that THR programs can considerably improve the social interaction and communication skills of children with ASD. Although we included only five articles in this study, they are most good quality and provide scientific evidence for our recommendation based on our quality assessment (see Table 2). All of the articles in our analysis described the use of horse-assisted therapy in their intervention groups focused on core ASD social and communication behaviors. We did not include studies of interventions involving the assistance of other animals (e.g., dog, cat, or dolphin) in our analysis, nor other study types, such as case studies, observational studies, or review articles. We calculated the ES of the studies, and we excluded studies with insufficient information for calculating the ES and those with low ES. By applying the above inclusion/exclusion criteria, we ensured that the quality of our study was well-controlled and focused only on THR interventions in relation to the social and communication skills of children with ASD.
The results of the present meta-analysis demonstrate the considerable effects of THR programs on core behaviors of children with ASD, in line the findings of current AATrelated studies, which showed how animal assistance can promote social interaction and communication skills [11,12,38]. Our results also reveal effects upon social behaviors, as evaluated through the Social Responsiveness Scale (SRS), which includes subscales for social communication, social awareness, social cognition, and social motivation. All of these showed significant improvements overall compared with the waiting list control group. We think that these improvements in core behaviors of ASD are because wellorganized THR programs provide opportunities for children with ASD to interact and communicate with horses, same-age peers, and instructors. In the five studies included, researchers provided evidence that improvements in social interaction and communication skills occurred because of the opportunities and stimulations gained from both verbal and nonverbal interactions with horses, peers, and trainers. Such interactions help children with ASD to better understand others, which is the most important means by which they can improve their social and communication skills.
The pooled meta-analysis results from two and three studies showed that no significant improvement was found concerning irritability or stereotypy behaviors in children with ASD. Our analysis is that these two areas of impairments are typical core behaviors of individuals with ASD and require careful, long-term remediation. The presentation of irritability behavior is usually associated with many other behavioral and emotional factors, including anxiety and mood difficulties, and these might trigger irritability behaviors. Stereotypy behavior is another impairment that needs long-term rehabilitation for correction because irritability and stereotypy behaviors in ASD may be attributed to neurobiological mechanisms, including dysfunction of the serotonergic, dopaminergic, and GABA neurotransmitter systems [47,48]. More specially, the basal ganglia are prone to dysfunction of the neurotransmitter system, which has been documented to contribute to stereotypical behaviors in ASD [49,50]. However, the changes in neurotransmitter systems require longer periods of exercise intervention (≥36 weeks) based on an animal model study [51]. In the studies included in our analysis, the experimental period was around 10 weeks, and a 10-week experimental period might be too short to reveal meaningful changes in the irritability and stereotypy behaviors of children with ASD. Our findings suggest this as a reasonable explanation for why no significant influence on irritability and stereotypy behaviors was observed.
THR programs are associated with the experience of happiness that comes from being in contact with nature and animals, and, as such, programs are normally designed to integrate many specific rehabilitation activities into the remediating objectives related to the children's behavioral problems; we think this is why THR programs are effective, enjoyable, and attractive to children with ASD. Although children with ASD may continue to demonstrate autistic behaviors such as irritability and stereotypy after undergoing a THR program, they typically show considerable improvements in other behaviors, such as social communication, social awareness, social cognition, and social motivation. Overall, such changes in behavior can produce many positive influences on the core social impairment of children with ASD.

Conclusions and Limitations
In this study, we used the techniques of a systematic review and meta-analysis to analyze existing research findings regarding the effects of THR programs on the social interaction and communication skill of children with ASD. The findings generated from the selected studies provided us evidence used to conclude that THR programs can notably remediate autistic impairments in social communication, social awareness, social cognition, and social motivation. This is in line with the findings of previous studies. The nature of THR programs, involving experiences of happiness, and being in contact with nature and animals, can considerably decrease levels of specific maladaptive behaviors of children with ASD, such as lethargy, hypersensitivity, and inappropriate speech. However, the findings from this meta-analysis do not provide statistical evidence of any influence upon the autistic behaviors of irritability and stereotypy, and we think that the short experimental periods used might have been an influencing factor in this regard. We also recognize that the findings derived from the existing RCT studies might be affected by the limited studies included in the present study, even though we followed a statistical meta-analysis procedure to control the quality of the selected articles. In future studies, researchers may need to increase the number of studies used to prevent the possible statistical influences of the small number of studies.