DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY REVIEW

Effects of hippotherapy and therapeutic horseback riding on postural

control or balance in children with cerebral palsy: a meta-analysis
MONIKA ZADNIKAR 1 | ANDREJ KASTRIN 2

1 Centre for Education and Rehabilitation for Children and Adolescents with Special Needs (CIRIUS), Kamnik. 2 Institute of Medical Genetics, University Medical Centre Ljubljana,
Ljubljana, Slovenia.
Correspondence to Monika Zadnikar, Centre for Education and Rehabilitation for Children and Adolescents with Special Needs (CIRIUS), SI-1240 Kamnik, Slovenia.
E-mail: [email protected]

PUBLICATION DATA AIM This research review and meta-analysis presents an overview of the effects of hippotherapy
Accepted for publication 26th January 2011. and therapeutic horseback riding (THR) on postural control or balance in children with cerebral
Published online 24th March 2011. palsy (CP).
METHOD To synthesize previous research findings, a systematic review and meta-analysis were
ABBREVIATIONS undertaken. Relevant studies were identified by systematic searches of multiple online databases
THR Therapeutic horseback riding from the inception of the database through to May 2010. Studies were included if they fulfilled the
CP Cerebral palsy following criteria: (1) quantitative study design, (2) investigation of the effect of hippotherapy or
THR on postural control or balance, and (3) the study group comprised children and adults with
CP. The selected articles were rated for methodological quality. The treatment effect was coded as
a dichotomous outcome (positive effect or no effect) and quantified by odds ratio (OR). The pooled
treatment effect was calculated using a random-effects model. Meta-regression of the effect
size was performed against study covariates, including study size, publication date, and
methodological quality score.
RESULTS From 77 identified studies, 10 met the inclusion criteria. Two were excluded because
they did not include a comparison group. Therapy was found to be effective in 76 out of 84 chil-
dren with CP included in the intervention groups. The comparison groups comprised 89 children:
50 non-disabled and 39 with CP. A positive effect was shown in 21 of the children with CP in the
comparison group regardless of the activity undertaken (i.e. physiotherapy, occupational therapy,
sitting on a barrel or in an artificial saddle). The pooled effect size estimate was positive (OR 25.41,
95% CI 4.35, 148.53), demonstrating a statistically significant effectiveness of hippotherapy or THR
in children with CP (p<0.001). Meta-regression of study characteristics revealed no study-specific
factors.
INTERPRETATION The eight studies found that postural control and balance were improved during
hippotherapy and THR. Although the generalization of our findings may be restricted by the rela-
tively small sample size, the results clearly demonstrate that riding therapy is indicated to improve
postural control and balance in children with CP.

Cerebral palsy (CP) is a term used to define a spectrum of syn- ments, speech and language disorders, orthopaedic
dromes of posture and motor impairment that results from an complications, and epilepsy.5,6 One of the most significant
insult to the developing central nervous system.1 This syn- problems in children with CP is defective postural control.
drome includes the following components: aberrant control of Maintaining postural control, required for the performance of
movement and ⁄ or posture, early onset, and no recognizable activities of daily living, is often a major challenge for children
underlying progressive pathology.2 Disorders of movement with CP.7–10
and posture are caused by damage to the motor cortex. The Postural control is organized at two functional levels.8–10
consequences of chronic muscle imbalance and the resultant The first level consists of a direction-specific adjustment, when
deformities can cause increasing disability with age.3,4 In addi- the equilibrium of the body is endangered, or the generation
tion to postural and motor abnormalities, people with CP of direction-specific patterns of postural adjustment.9,10 For
may exhibit secondary consequences of brain damage, includ- example, when reaching, the muscles on the dorsal side of the
ing learning disability,* other cognitive and sensory impair- body are primarily activated when the body leans forwards,
whereas the muscles on the ventral side of the body are
*North American usage: mental retardation. primarily activated when the body leans in the opposite

684 DOI: 10.1111/j.1469-8749.2011.03951.x ª The Authors. Developmental Medicine & Child Neurology ª 2011 Mac Keith Press
direction.7,10 According to van der Heide the second level of What this paper adds
postural control is involved in the fine-tuning of the basic, • It is the first known meta-analysis of the literature on hippotherapy and thera-
direction-specific adjustment according to multisensorial peutic horseback riding in children with cerebral palsy.
afferent input from the somatosensory, visual, and vestibular • It gives an overview of the effects of hippotherapy and therapeutic horseback
systems. This modulation can be achieved in various ways, for riding on postural control or balance.
• The meta-analysis found a statistically significant effect of hippotherapy and
instance, by changing the order in which the agonist muscles therapeutic horseback riding in children with cerebral palsy.
are recruited (e.g. in a caudal to cranial sequence or vice versa),
by modifying the size of the muscle contraction, which is needed, but this should not be restrictive.20,22 The horse is
reflected in electromyography (EMG) amplitude, or by alter- equipped with a bridle and a vaulting surcingle, which can be
ing the degree of antagonist activation.10 used as a handhold and to which auxiliary reins are attached.
Various approaches exist for improving postural control The exercises are aimed primarily at maintaining balance and
and balance, including neurodevelopmental treatment which proper body posture in different positions, at the development
is classified as a neurophysiological or neuromaturational ther- of the rider’s sensory–motor and perceptual–motor skills and
apeutic approach, as are Vojta therapy, Temple Fay, the Rood at a gradual increase in the rider’s capacity to stretch and move
and Kabat methods, and sensory integration.9–11 Newer while the horse moves at a slow, steady gait.20,22
approaches, such as the ecological and dynamical system The primary goal of hippotherapy is to improve the individ-
approaches, propose that development of motor skills and ual’s balance, posture, function, and mobility. Hippotherapy is
coordination occurs as the result of multifaceted interactions an individualized treatment that uses an interdisciplinary team
that take place in the context of performing a specific task.10,11 approach.23,24 The treatment is administered by a trained
Other frequently evaluated therapeutic methods are conduc- health professional (physiotherapist, occupational therapist, or
tive education and biofeedback with a central role for motor speech therapist). Hippotherapy has been used for over
learning theories.11 Hippotherapy is also one of these 30 years in the treatment of children with spastic CP. During
approaches in which equine movement is used for its thera- hippotherapy, the therapist focuses on improving walking abil-
peutic effect.12 ity, posture, balance, tone, and mobility.16,25 The literature on
hippotherapy research discusses its physical and psychological
Therapeutic horseback riding and hippotherapy benefits. The reported physical benefits include improvement
Riding may lead to improved coordination,13 increased head in balance, strength, coordination, muscle tone, joint range of
and trunk control,14 and improved gait.15,16 The rationale for movements, weight-bearing, posture, gait, and sensory pro-
hippotherapy is that the horse’s gait provides a precise, cessing.17,26,27 The psychological effects are seen in improved
smooth, rhythmic, and repetitive pattern of movement to the self-confidence, self-esteem, motivation, attention span, spatial
rider that is similar to the mechanics of human gait.15,17 The awareness, concentration, and verbal skills.25,26,28
horse’s centre of gravity is displaced three-dimensionally when An understanding of the current applications of hippotherapy
walking, resulting in a movement that is very similar to that of and THR in the population with CP could serve as a basis for
the human pelvis during walking.17 This rhythmical move- further research, assist in clinical management, and improve
ment, combined with the warmth of the horse, is hypothesized the quality of life of children and adults with the disorder.
to decrease hypertonicity and promote relaxation in the rider Recent systematic reviews22,29,30 that included several clinical
with spastic CP.18 Adjusting to the horse’s movements also trials, established a positive effect of hippotherapy and THR.
involves the use of muscles and joint movements which, over Because CP is integrally related to postural control and bal-
time, may lead to increased strength and range of motion.12,19 ance, it seems important to evaluate the effectiveness of riding
In general, the movement of the horse provides a variety of therapy by examining these parameters. Currently, there is no
inputs to the rider, which may be used to facilitate improved meta-analysis of the effectiveness of hippotherapy and THR
contraction, joint stability, weight shift, and postural equilib- on postural control and balance in the population with CP.
rium responses in children with CP.6 Hence, the aim of this meta-analysis was to summarize and
Therapeutic horseback riding (THR) is a broad term that evaluate critically the evidence for or against the effectiveness
covers many elements of recreational horse riding. THR must of hippotherapy and THR in people with CP.
be performed only by specially trained riding instructors and
assistants. To ensure safety and effectiveness, the instructors, METHOD
even though they are usually not medical professionals, should This meta-analysis was completed in accordance with
be aware of the rider’s state of health, contraindications, dis- Preferred Reporting Items for Systematic Reviews and Meta-
abilities, and other limitations, and be competent to select, Analyses (PRISMA) guidelines (Table SI, supporting material
train, and prepare the appropriate horse.20,21 THR instructors published online).31
and assistants must follow multifaceted lesson plans as THR
sessions include many procedures and precautions. The THR Search strategy
instructor decides if a saddle is required and if the horse A systematic literature search of a total of 11 medical, scien-
remains stationary or walks, while the child, following his tific, and arts bibliographical databases was performed to
directions, tries, for example, to touch different parts of the identify citations relevant to the effectiveness of hippotherapy
horse or reach for an object. The assistant provides help if and THR for children with CP. The search strategy com-

Review 685
prised searches of the following electronic databases: Web of Data extraction
Science, MEDLINE (through PubMed), ProQuest, Current All articles were read by two investigators who independently
Contents, The Cochrane Library, Cochrane Database of extracted data from the articles. The following information
Systematic Reviews, the Cochrane Controlled Trials Regis- was sought from each article: author identification, year of
ters, Ovid, Embase, CINHAL, and Google Scholar. The publication, type of study design, study population, sample
resulting search terms were combined as follows: ‘develop- size, and results. The treatment effect on postural control or
mental riding therapy’ OR ‘equine-movement therapy’ OR balance was coded as a dichotomous outcome (positive effect
‘riding therapy’ OR ‘riding for disabled’ OR ‘therapeutic or no effect). Disagreements were resolved by consensus
horseback riding’ OR ‘therapeutic riding’ OR ‘hippotherapy’ between the investigators.
AND ‘cerebral palsy’ OR ‘posture control’ OR ‘balance’. (For
details on search strategy for MEDLINE, see Table SII, Statistical analysis
supporting material published online.) Individual strategies Study-specific treatment effects were quantified by odds ratio
were developed for each source searched to accommodate idi- (OR) with 95% confidence intervals (CI). Because the OR is
osyncrasies of search engines. The search dates covered the strictly positive, we chose to analyse the natural logarithm of
period from the inception of the databases until 20 February the OR. For individual studies with no events in one or both
2010. The search was updated in May 2010, without the groups, a continuity correction of 0.5 was used. Unless other-
addition of further data. Reference lists of identified articles wise stated, a p-value of less than 0.05 was considered to indi-
and the reference lists of previous reviews22,29,30 were also cate statistical significance of the derived results. For
manually scanned for possible relevant articles not previously calculation of the combined treatment effects, the fixed-effects
identified. model (Mantel–Haenszel method) and the random-effects
model (DerSimonian–Laird method) were used. The selection
of a random- or fixed-effects model in meta-analysis is contro-
Selection of studies for inclusion
versial. The random-effects model for pooling effects was pre-
Studies were selected for inclusion if they fulfilled the follow-
ferred in cases of heterogeneity of treatment effect. Statistical
ing criteria: (1) a quantitative study design; (2) the investiga-
heterogeneity was assessed by means of a Mantel–Haenszel
tion of the effect of hippotherapy or THR on postural
derived Cochran’s Q statistic. Cochran’s Q was used to test the
control or balance; and (3) the study group comprised chil-
null hypothesis that all treatment effects are equivalent. If p
dren and adults with CP. To increase the reliability of cita-
was less than 0.10, the heterogeneity was considered statisti-
tion selection, all potentially relevant citations were reviewed
cally significant, and the random-effects model was then used.
independently by two investigators. A consensus was reached
Heterogeneity was also quantified using the I2 metric, which is
between the researchers about whether the article fulfilled
independent of the number of studies in the meta-analysis
the inclusion criteria. Full copies of all selected articles were
(I2<25%, no heterogeneity; I2=25–50%, moderate heterogene-
retrieved. Reviews, correspondence, and editorials were spe-
ity; I2>50%, large or extreme heterogeneity).33 Several meth-
cifically excluded, although their reference lists were scanned
ods were used to assess the potential for publication bias.
to identify possible relevant studies. We attempted to contact
Visual inspection of the funnel plot was conducted. A funnel
the corresponding author for clarification of data extraction
plot allows evaluation of publication bias by presenting the
or quality assessment if this was not clear from the published
study’s log OR as a function of its standard error. The Begg
article or abstract. Two investigators then independently
rank correlation method was also used to assess publication
reviewed the full articles to determine whether they indeed
bias formally. Standard meta-regression of the effect size
met the inclusion criteria. Disagreements were resolved by
expressed as log OR was performed against study covariates,
discussion.
including study size, publication date, and the Critical Review
Form score. Meta-regression was weighted by the inverse vari-
Assessment of study quality ance of each study. Data management, statistical analysis, and
The methodological quality of each study was evaluated inde- graphic visualization were performed using the R program-
pendently by two reviewers using the Critical Review Form – ming language (http://www.r-project.org) with the rmeta and
Quantitative Studies.32 Results were compared and any dis- meta packages. The complete R ⁄ Sweave source code to repro-
agreement resolved by discussion. The following main catego- duce the results of this meta-analysis is available in Figure S1
ries were considered: (1) study purpose, (2) literature review, (supporting material published online).
(3) study design, (4) sample, (5) outcomes, (6) interventions,
(7) results, (8) conclusion, and (9) clinical implications. The RESULTS
overall quality of each article was examined using 16 dichoto- The study selection flow diagram is shown in Figure 1. Our
mous items evaluating the internal and external validity of the searches identified 77 potentially relevant studies. Eight stud-
study, its findings, and conclusions. These 16 questions were ies fulfilled all inclusion criteria, and there was 100% agree-
scored as either 1 (completely fulfils the criterion) or 0 (does ment between the investigators regarding article eligibility.
not fulfil the criterion) and the scores totalled for each study. Two studies were excluded because the comparison group was
A maximum score of 16 indicated excellent methodological not clearly defined.34,35 All identified articles were published
quality. between October 1988 and December 2009 in peer-reviewed

686 Developmental Medicine & Child Neurology 2011, 53: 684–691

 77 records identified statistically significant association was found in only three
through database studies (data not shown; Bertoti,17 Kuczynski and Słonka,36
searching Shurtleff et al.14). The DerSimonian–Laird estimate of pooled
OR was 25.41, 95% CI 4.35, 148.53, z=3.59, p<0.001. Pooled-
effect size estimate was positive, demonstrating a statistically
71 records after significant effectiveness of hippotherapy or THR in children
duplicates removed
with CP. There was distinct heterogeneity among the studies
(Q(7)=17.81, p=0.01; I2=60.7%). Figure 2 shows a forest plot of
the random-effects model ordered by date of publication with
71 records screened 4 records excluded OR and corresponding CI. Meta-regression analysis of study
characteristics (study size, publication date, study quality
score) found no study-specific factors (R2adj =0.51, F(3,4)=3.44,
10 full-text articles 2 full-text articles p=0.132).
assessed for eligibility excluded
Figure 3 shows a funnel plot for the visual inspection of
publication bias. The plot shows slight asymmetry, indicating
a possible minor publication bias in favour of the null associa-
8 studies included in
tion. However, Begg’s adjusted rank correlation test indicated
qualitative synthesis
no evidence of a substantial publication bias (z=1.48, p=0.138).

8 studies included in DISCUSSION

quantitative synthesis Knowledge of equine-assisted therapies, especially hippotherapy
(meta-analysis) and THR, and their efficacy is continually increasing and is
becoming more reliable and precise. We have performed the
first known meta-analysis of the literature on hippotherapy or
Figure 1: The outline of the search strategy. THR on patients with CP. Using pooled data from eight
studies, we have found a statistically significant effectiveness of
journals. Quality scoring of these studies according to the hippotherapy or THR among children with CP.
Critical Review Form – Quantitative Studies scale was under- Postural control and balance are important elements of
taken by the same investigators, with full final agreement motor tasks. Postural stability is essential for the accomplish-
(Tables I and SIV, supporting material published online). The ment of increasingly complex motor skills and coordination.
median score was 10.5. Characteristics of the studies included Several sensory deficits associated with CP, including prob-
in the meta-analysis are shown in Table II. The eight studies lems with vision, proprioception, and cutaneous perception,
included 84 children with CP in the intervention groups, 76 of may contribute to impaired postural and balance control.
whom felt positive effects. The comparison groups comprised Hippotherapy and THR are comprehensive strategies to
89 children, including 39 with CP. Twenty-one of the chil- improve postural control and balance in children with CP,
dren with CP felt positive effects from continuation of their and consequently daily activities, independence, and quality of
weekly physiotherapy and ⁄ or occupational therapy, or from life, as shown in all eight studies. A significant improvement
sitting on a barrel or in an artificial saddle. The remaining 50 was documented in studies by Benda et al.,3 Bertoti,17 Kuczyn-
children were non-disabled, and no effect on postural control ski and Słonka,36 Quint and Toomey,19 and Shurtleff et al.,14
and balance was found in this subgroup. by MacKinnon et al.28 in children with moderate CP, and by
The meta-analysis was performed on the overall sample of MacPhail et al.37 in children with diplegic CP. However, we
the selected clinical studies. All the studies showed a positive were unable to find any significant improvement in the study
effect after treatment with hippotherapy or THR, although a by Haehl et al.,27 although their conclusion was that hippo-
therapy has a very positive effect on children with CP, includ-
ing improvement in postural control and balance.
Table I: Methodological quality of research articles In three of the studies3,14,27 the effects were measured
after hippotherapy and in three after THR.17,28,37 Two of
Study Quality score the studies demonstrated the effects of measurements after
sitting in an artificial Brunel active balance saddle with
Benda et al.3 13
Bertoti17 13 precise constant amplitude of movement.19,36 All the studies
Haehl et al.27 6 established the suitability of hippotherapy and THR for
Kuczynski and Słonka36 10
children with CP. Both types of riding and Brunel active
MacKinnon et al.28 5
MacPhail et al.37 11 balance saddle improved postural control and balance.
Quint and Toomey19 7 During hippotherapy or THR, changes in the horse’s gait
Shurtleff et al.14 13
velocity facilitate righting and equilibrium responses, which
Median value 10.5
enable the development of dynamic postural stabilization
The highest possible score is 16. and postural control.22

Review 687
 Table II: Summary of articles on hippotherapy, therapeutic horseback riding (THR), and the Brunel active balance saddle (BABS)

Effect on postural
Sampling (structure of the intervention Intervention duration of control or balance
Study and comparison groups: number, age, therapy performed in IntG and
Study design CP, or ND) CompG Measures, tests Outcome, results IntG CompG

Bertoti17 PED IG: 11; 2–9y; 10wk no riding Two pretests Pretest 1 median 20 11+ 4+
CP CG: 11; 10wk THR One posttest Pretest 2 median 22; 7)
2–9y; CP Twice weekly, 60min Posture assessment scale: Posttest: median 27 Q 12.86,
IntG riding Bertotti test p£0.05
CompG no riding Friedman Reliability r=0.82;
Postural controla
MacKinnon et al.28 RCT IG: 10; 5–11y; 26wk Bertoti test PDMS Aa 5+ 4+
CP CG: 9; Once weekly, 60min GMFM CBCL activitiesa 5) 5)
4–9y; CP IntG: THR PDMS Mild CP
CompG: no riding BOTMP Moderate CPa
VABS
SPPC
CBCL

688 Developmental Medicine & Child Neurology 2011, 53: 684–691

Video
MacPhail et al.37 QED IG: 6; 5–12y; One intervention Video in the frontal plane: Lateral trunk deviation: ND 3+ 7)
CP CG: 7; IntG: THR lateral deviation trunk and 5.8; SD 0.5 CP 10.2, SD 2.2; 3)
6–11y; ND CompG: riding pelvis of the rider and p<0.001
horse CP diplegiaa
t-test CP tetraparesis

Quint and Toomey19 RCT IG: 13; 9–16y; 4wk Passive range of Passive range of 13+ 9+
CP CG –13; Sitting on BABS or static antero-posterior pelvic tilt antero-posterior pelvic tilt 4)
9–16y CP saddle Pretest ⁄ posttest with BABS increase by 20.8a
10·10min photography Static saddle increase by 8.85
IntG: BABS t-test t-test 2.654
CompG: sitting on static saddle
Haehl et al.27 QED IntG: 2; 9 and 4y; 12wk HT First phase: video ND Postural control 2+ 2)
CP CompG: 2; 9 Once weekly Second phase: video CP, Postural coordination
and 7y; ND One child 20min test postural control and PEDI 1a, 1
One child 40min postural coordination
IntG.: HT Pretest ⁄ posttest PEDI
CompG: one riding session
Kuczynski and Słonka36 QED IntG: 25; 3–10y; 12wk Pretest ⁄ Posttest Postural control 25+ 33)
CP CompG: 33; Twice weekly, 20min Stabilography: measure the Pretest: sagit 11.2–5.7a, front
3–10y; ND IntG: BABS excursion of centre of 11.5–6.4
CompG: one session on BABS, pressure Posttest: sagit 9.0–4.1a, front
pre ⁄ posttests Autoregressive modelling 8.1–4.0 FRD (feet): p>0.07

Benda et al.3 RCT IntG: 7; 4–12y; One intervention Pretest ⁄ posttest Mean asymmetry: 7+ 4+
CP CompG: 6; IntG: 8min HT, once CompG: Muscle symmetry: EMG 55.5 (SD 82.5), HT 11.9 (SD 2)
4–12y; CP 8min sitting on barrel, once 10s sitting, 10s standing, 29.9), barrel t(11)=1.22;
IntG: HT 10 feet walking p=0.24a;
CompG: sitting on barrel Video Mean (%):
64.6 (SD 28.3) HT )12.8 (88.8)
barrel
t(11)=2.19; p=0.051
 Different methods were used to measure the effectiveness of

CompG
control or balance
Effect on postural
hippotherapy and THR. These were supported by various

Developmental Motor Scales; CBCL, Child Behavior Checklist; BOTMP, Bruininks-Oseretsky Test of Motor Proficiency; VABS, Vineland Adaptive Behavior Scales; SPPC, Self-Perception Profile for
Children; HT, hippotherapy; PEDI, Pediatric Evaluation of Disability Inventory; FRD, feet-related displacement; CP, cerebral palsy; ND, non-disabled; IntG, intervention group; CompG, comparison
8)
tests, which were conducted before and after riding. In three
studies,14,27,37 results were obtained from video recordings. In
the study by Benda et al.,3 the results were measured using

IntG

Statistically significant result. PED, pre-experimental design; RCT, randomized control trial; QED, quasi-experimental design;38 GMFM, Gross Motor Function Measure; PDMS, Peabody
10+
surface EMG. Kuczynski and Słonka36 measured postural bal-
ance by centre-of-pressure displacement using a force plate.
Quint et al.19 used photographs to assess passive range of

Functional reach test; p<0.05;

Dynamic trunk ⁄ head; p<0.05;

Reliability r=0.95 post 1 and

movement of the pelvis. Many additional tests, which were
post 1 and post 2a themselves the basis of some of the included studies, also

post 1 and post 2a

showed changes (Table II).
Outcome, results

In four out of eight studies, children with CP were included

r=0.94 post 2 in the comparison group,3,17,19,28 whereas in four studies the
comparison group comprised non-disabled children.14,27,36,37
No changes in postural control or balance were recognized in
non-disabled children, but an effect on postural control and
balance was observed in children with CP in the comparison
group. In 21 of 39 children with CP, who continued with their
Second posttest after
Functional reach test
dynamic trunk ⁄ head
stability: barrel test

regular physiotherapy and ⁄ or occupational therapy during the

Pretest ⁄ posttest
Measures, tests

study period, or sat on a barrel or in an artificial saddle, a posi-

tive effect was observed. In one study the same children were
12–14wk

included in both the intervention and comparison groups,

Video

with riding therapy being undertaken during inclusion in the

intervention group.17 In two studies, the comparison group
group; +, effect on postural control or balance; ), no effect on postural control or balance. Q, Friedman’s Q statistic.

also received a single riding session.14,36 Two comparison

CompG: one session on barrel,

groups did not have any additional tasks; hence children with
therapy performed in IntG and

CP continued with the same therapy schedule they had

barrel test, and pre ⁄ post

followed before inclusion in the study.17,28 One comparison

Intervention duration of

functional reach test

group sat on a barrel,3 and two groups on the Brunel active

Once weekly, 45min

balance saddle19,36 (Table II). In two of the studies, the chil-

dren received riding therapy only once, with measurements
being performed before and after riding,3,37 whereas in other
IntG: HT
CompG

studies the riding intervention continued for 4 to 12 weeks,

12wk

and for 6 months in the study by MacKinnon et al.28

All the authors concluded that there are positive effects
on postural control or balance after hippotherapy, THR, or
Sampling (structure of the intervention
and comparison groups: number, age,

sitting in an artificial saddle. Nevertheless, these studies

have some limitations. Differences in effects were observed
between children with mild and moderate CP.27,28 None of
the selected studies included adults with CP. The sample
size of children with CP in the intervention group was rela-
tively small, and the children also differed in the type and
severity of CP (mild and moderate, and diplegia, tetraplegia,
IntG: 10; 5–13y;
CP CompG: 8;

and hemiplegia). The comparison group comprised children

5–13y; ND
CP, or ND)

with CP and non-disabled children, which makes compari-

son of results extremely difficult. The authors also con-
cluded that the number of children with disabilities in these
groups (intervention and comparison) was too small to gen-
eralize the conclusion to the entire population of people
design
Study

with CP. In addition, the activities themselves differed

QED

between the studies. In all intervention groups children

Table II: (Continued).

were treated with hippotherapy and THR. Moreover, all

the eight studies used different research methods and tests,
Shurtleff et al.14

the instruments for measuring the effects were very diverse,

and the duration of therapy was extremely variable, ranging
Study

from a single measurement to 6 months of measurements.

Precisely because of the diversity of results, we decided to
a

Review 689
 Experimental Control
Study Events Total Events Total OR 95%CI W(fixed) W(random)

Bertoti17 11.5 12 4.5 12 38.33 [ 1.79, 820.13] 5.7% 13.2%

MacKinnon et al.28 5.0 10 4.0 9 1.25 [ 0.21, 7.62] 64.1% 17.9%
MacPhail et al.37 3.5 7 0.5 8 15.00 [ 0.60, 376.70] 7.1% 12.7%
Quint and Toomey19 13.5 14 9.5 14 12.79 [ 0.61, 266.66] 10.3% 13.3%
Haehl27 2.5 3 0.5 3 25.00 [ 0.34, 1831.59] 2.5% 9.6%
Kuczynski et al.36 25.5 26 0.5 34 3417.00 [65.56, 178107.63] 0.3% 10.5%
Benda et al.3 7.5 8 4.5 7 8.33 [ 0.32, 215.68] 9.1% 12.6%
Shurtleff and Słonka14 10.5 11 0.5 9 357.00 [ 6.39, 19942.80] 0.8% 10.3%

Fixed effect model 91 96 18.17 [ 7.80, 42.32] 100% −−

Random effects model 25.41 [ 4.35, 148.53] −− 100%

0.75
Odds Ratio

Figure 2: Forest plot. For each study, the point estimate is given by a square whose size is inversely proportional to the standard error of the estimate. The
95% confidence interval for each study is given by a horizontal line around the estimate. The summary odds ratio is drawn as a diamond with horizontal limits
at the confidence limits and width inversely proportional to its standard error.

This meta-analysis and the research review provide the basis

0.0

for a future research article on the implementation and effec-

tiveness of hippotherapy and THR in children and adults with
CP, not only for postural control and balance, but also for
0.5

other motor functions that are of extreme importance to peo-

ple with CP. Additional research using a larger sample size
and matched comparisons is needed to investigate the effec-
Standard error

MacKinnon tiveness of hippotherapy and THR on gait, joint stability,

1.0

co-contraction, gross motor function, movement of the pelvis,

other functions, and changes in physical and psychosocial inter-
action. Blinded assessment and non-riding controls are obliga-
1.5

Quint Bertoti tory. In addition to the short-term effects of hippotherapy and

Benda MacPhail THR, which are measured immediately after the riding ses-
sion, the long-term clinical effects should be measured one or
Kuczynski two years after completion of the study. Adults with CP (in
2.0

Shurtleff
whom growth and development have been completed) should
Haehl
also be included in future studies to establish the effectiveness
−2 0 2 4 6 8 of hippotherapy and THR on their motor function. We are of
Log odds ratio the opinion that it is unrealistic to expect that research proto-
cols and data analysis will be standardized. Similarly, it will be
difficult to obtain a large enough sample size and sufficient
Figure 3: Funnel plot with the pooled estimate of the random-effects
numbers of children with different forms of CP (diplegia,
model. Diagonal lines define a region within which 95% of points might lie
hemiplegia, tetraplegia).
in the absence of publication bias.

CONCLUSION
In this review article all the research data on the influence of
categorize the studies according to whether the effects of hippotherapy and THR on postural control or balance, on
hippotherapy and THR on postural control and balance intervention and comparison groups, length of each session
were positive or there was no effect. and duration of hippotherapy and THR, and tests used as the
As with any meta-analysis, the potential for publication bias main source for proof of the effects or measurements are col-
is a major concern. Publication bias results from selective lected. What becomes evident throughout the period of dem-
reporting of studies featuring positive results, potentially lead- onstration and measurement of the effectiveness of both
ing to the over-representation of the benefit of hippotherapy hippotherapy and THR is that the intervention and compari-
or THR in published papers. Our funnel plot appears slightly son sample sizes are too small, and the population with CP
asymmetrical, meaning that there may be publication bias, extremely diverse. This is reflected in the complexity of man-
which is difficult to quantify. However, publication bias was agement of this population regardless of the form of treat-
not detected when the Begg rank correlation method was ment, not only hippotherapy and THR. This research
applied. evidence and meta-analysis suggest that clinicians and thera-

690 Developmental Medicine & Child Neurology 2011, 53: 684–691

pists can recommend either hippotherapy or THR as forms of ACKNOWLEDGEMENTS
therapy to improve posture and balance, and consequently to We thank Snežana Bitenc for her assistance in cross-referencing the
influence functioning in activities of daily life and quality of literature, Anastasija Lavrič for independent verification and review of
life in children with CP. all treatment effects in selected articles, and Dianne Jones for lan-
guage review. The second author was supported by the Slovenian
ONLINE MATERIAL ⁄ SUPPORTING INFORMATION Research Agency.
Additional material and supporting information for this paper may be
found online.

REFERENCES
1. Koman LA, Smith BP, Shilt JS. Cerebral palsy. Lancet 2004; 14. Shurtleff TL, Standeven JW, Engsberg JR. Changes in 27. Haehl V, Giuliani C, Lewis C. Influence of hippotherapy on
363: 1619–31. dynamic trunk ⁄ head stability and functional reach after the kinematics and functional performance of two children
2. Taylor JR, Kopriva PG. Cerebral palsy. In: Brodwin MG, hippotherapy. Arch Phys Med Rehabil 2009; 90: 1185–95. with cerebral palsy. Pediatr Phys Ther 1999; 11: 89–101.
Tellez FA, Brodwin SK, editors. Medical, Psychosocial, and 15. McGee MC, Reese NB. Immediate effects of a hippotherapy 28. MacKinnon RJ, Noh S, Lariviere J, MacPhail A, Allan DE,
Vocational Aspects of Disability. Athens: Elliott Fitzpatrick, session on gait parameters in children with spastic cerebral Laliberte D. A study of therapeutic effects of horseback rid-
2002: 387–400. palsy. Pediatr Phys Ther 2009; 21: 212–8. ing for children with cerebral palsy. Phys Occup Ther Pediatr
3. Benda W, McGibbon NH, Grant KL. Improvements in 16. Winchester P, Kendall K, Peters H, Sears N, Winkley T. 1995; 15: 17–34.
muscle symmetry in children with cerebral palsy after The effect of therapeutic horseback riding on gross motor 29. Snider L, Korner-Bitensky N, Kammann C, Warner S, Saleh
equine-assisted therapy (hippotherapy). J Altern Complement function and gait speed in children who are developmentally M. Horseback riding as therapy for children with cerebral
Med 2003; 9: 817–25. delayed. Phys Occup Ther Pediatr 2002; 22: 37–50. palsy: is there evidence of its effectiveness? Phys Occup Ther
4. Massion J, Alexandrov A, Frolov A. Why and how are pos- 17. Bertoti DB. Effect of therapeutic horseback riding on posture Pediatr 2007; 27: 5–23.
ture and movement coordinated? Prog Brain Res 2004; 143: in children with cerebral palsy. Phys Ther 1988; 68: 1505–12. 30. Debuse D, Chandler C, Gibb C. An exploration of German
13–27. 18. Freeman G. Hippotherapy: therapeutic horseback riding. and British physiotherapists’ views on the effects of hippo-
5. Best SJ. Cerebral Palsy. Teaching Individuals with Physical Clin Manag Phys Ther 1984; 4: 20–5. therapy and their measurement. Physiother Theory Pract
or Multiple Disabilities. 5th edn. Upper Saddle River: Pre- 19. Quint C, Toomey M. Powered saddle and pelvic mobility: 2005; 21: 219–42.
ntice Hall, 2005. an investigation into the effects on pelvic mobility of children 31. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred
6. Miller F. Physical Therapy of Cerebral Palsy. New York: with cerebral palsy of a powered saddle which imitates the reporting items for systematic reviews and meta-analyses: the
Springer, 2007. movements of a walking horse. Physiotherapy 1998; 84: 376– PRISMA statement. PLoS Med 2009; 6: e1000097.
7. van der Heide JC, Begeer C, Fock JM, et al. Postural control 84. 32. Law M, Stewart D, Pollicj N, Letts L, Bosch J, Westmorland
during reaching in preterm children with cerebral palsy. Dev 20. North American Riding for the Handicapped Association. M. Critical Review Form – Quantitative Studies; and Guide-
Med Child Neurol 2004; 46: 253–66. Annual Report. Denver, CO: North American Riding for the lines for Critical Review Form – Quantitative Studies. Ham-
8. Forssberg H, Hirschfeld H. Postural adjustments in sitting Handicapped Association, 2004. ilton, Ontario: McMaster University, 1998.
humans following external perturbations: muscle activity and 21. Sterba JA, Rogers BT, France AP, Vokes DA. Horseback rid- 33. Higgins JP, Thompson SG. Quantifying heterogeneity in a
kinematics. Exp Brain Res 1994; 97: 515–27. ing in children with cerebral palsy: effect on gross motor meta-analysis. Stat Med 2002; 21: 1539–58.
9. Shummway-Cook A, Woollacott MH. Motor Control: function. Dev Med Child Neurol 2002; 44: 301–8. 34. Hamill D, Washington K, White OR. The effect of hippo-
Theory and Practical Applications. 2nd edn. Philadelphia: 22. Sterba JA. Does horseback riding therapy or therapist-direc- therapy on postural control in sitting for children with cere-
Lippincott Williams & Wilkins, 2001. ted hippotherapy rehabilitate children with cerebral palsy? bral palsy. Phys Occup Ther Pediatr 2007; 27: 23–42.
10. van der Heide JC, Hadders-Algra M. Postural muscle dysco- Dev Med Child Neurol 2007; 49: 68–73. 35. Land G, Errington-Povalac E, Paul S. The effects of thera-
ordination in children with cerebral palsy. Neural Plast 2005; 23. All AC, Loving GL, Crane LL. Animals, horseback riding, peutic riding on sitting posture in individuals with disabili-
12: 197–203. and implications for rehabilitation therapy. J Rehabil 1999; ties. Occup Ther Health Care 2002; 14: 1–12.
11. Siebes RC, Wijnroks L, Vermeer A. Qualitative analysis of 65: 49–57. 36. Kuczyński M, Słonka K. Influence of artificial saddle riding
therapeutic motor intervention programmes for children 24. Strauß I. Hippotherapie. 3th edn. Stuttgart: Hipokrates, on postural stability in children with cerebral palsy. Gait Pos-
with cerebral palsy: an update. Dev Med Child Neurol 2002; 2000. ture 1999; 10: 154–60.
44: 593–603. 25. McGibbon NH, Andrade CK, Widener G, Cintas HL. 37. MacPhail A, Edwards J, Golding J, Miller K, Moiser C,
12. McGibbon NH, Benda W, Duncan BR, Silkwood-Sherer D. Effect of an equine-movement therapy program on gait, Zwiers T. Trunk postural reactions in children with and
Immediate and long-term effects of hippotherapy on symme- energy expenditure, and motor function in children with without cerebral palsy during therapeutic horseback riding.
try of adductor muscle activity and functional ability in chil- spastic cerebral palsy: a pilot study. Dev Med Child Neurol Pediatr Phys Ther 1998; 10: 143–7.
dren with spastic cerebral palsy. Arch Phys Med Rehabil 2009; 1998; 40: 754–62. 38. DePoy E, Gitlin LN. Introduction to Research: Understand-
90: 966–74. 26. Murphy D, Kahn-D’Angelo L, Gleason J. The effect of ing and Applying Multiple Strategies. 2nd edn. Chicago:
13. Potter JT, Evans JW, Nolt BH. Therapeutic horseback rid- hippotherapy on functional outcomes for children with Mosby, 1998.
ing. J Am Vet Med Assoc 1994; 204: 131–3. disabilities: a pilot study. Pediatr Phys Ther 2008; 20: 264–70.

Review 691