Equine-Assisted Therapy

Abstract

Equine-assisted therapy is a type of animal assisted therapy in which a horse is an integral, planned part of a documented treatment delivered by a credentialed professional. The horse supplies two things at once no other therapy animal does: a rhythmic, three-dimensional gait that drives the rider's pelvis close to the pattern of human walking, and a prey animal's acute sensitivity to a handler's arousal that makes it an immediate behavioural mirror. From these follow its two application families, the sensorimotor work of hippotherapy and the relational work of equine-assisted psychotherapy. The evidence is contested: small trials report moderate benefits for autism, psychiatric, and trauma populations, while skeptics argue the effects lack the dose and sample sizes needed to be believed. Three demos trace the gait oscillation, the arousal mirror, and how a moderate effect size survives a realistic sample.

Keywords: equine-assisted therapy, hippotherapy, human-animal bond

What Equine-Assisted Therapy Is

Equine-assisted therapy is the deliberate inclusion of a horse in a structured treatment plan, where the animal's movement and behaviour are instruments toward defined clinical goals. It is the sole narrower descriptor MeSH files beneath animal assisted therapy, and it inherits that field's governing distinction: a goal, a plan, documented progress, and a credentialed professional are what separate a therapy from a pleasant encounter at a stable (#ref-iahaio-2018). The umbrella masks real internal variety, and a recurring problem in the literature is that the label equine-assisted therapy is stretched across interventions with incompatible mechanisms and aims (#ref-white-lewis-2020).

Four terms mark the field's live distinctions. Hippotherapy uses the horse's gait as a treatment tool, a physical or occupational therapist grading the movement to a rehabilitation goal. Therapeutic riding teaches riding skills to people with disabilities for their functional and recreational benefit. Equine-assisted psychotherapy places a mental-health professional and, usually, an equine specialist with a client and horse, working almost always from the ground rather than mounted, toward psychological goals. Equine-assisted learning applies the same ground-based format to educational and personal-development objectives rather than clinical ones. Conflating them is the single largest source of overstated and unpoolable claims (#ref-white-lewis-2020).

The horse’s walk as a movement prescription

At the walk a horse delivers a rhythmic, three-dimensional oscillation to the seated rider’s pelvis that closely mimics the pelvic motion of human walking. Scrub through one stride to trace the figure-eight the pelvis follows; set the cadence to see how many walking-like oscillations a session supplies.

This is the sensorimotor channel: a child who cannot walk still receives the graded, repetitive pelvic input of walking, which is why hippotherapy has its firmest evidence in gross-motor outcomes.

The therapeutic unit is larger than the client-animal dyad. In mounted work a physiotherapist, a horse leader, and often two sidewalkers surround the rider; in ground-based psychotherapy a mental-health clinician and an equine specialist share responsibility for the client and the horse. The horse is neither a passive apparatus nor a comfort object but a partner whose size makes safety a standing constraint and whose reactivity makes it an instrument (#ref-selby-2013).

Physiological and Affective Mechanisms

Two mechanisms run in parallel, and the horse is unusual in engaging both strongly. The relational mechanism is the one equine-assisted therapy shares with the wider human-animal field: calm, affiliative contact is accompanied by a rise in oxytocin and a fall in cortisol and blood pressure in both partners, the neuroendocrine signature that Beetz and colleagues synthesised into an oxytocin-centred model of stress buffering and social facilitation (#ref-beetz-2012). The interspecies oxytocin loop that Nagasawa and colleagues documented between dogs and owners — mutual attention raising the hormone in both, each response feeding the other — supplies the template for the bond a horse programme tries to recruit, even where the equine data themselves remain thinner than the canine (#ref-nagasawa-2015).

What the horse adds to the relational channel is amplification through prey-animal vigilance. A horse continuously reads posture, breathing, and movement for threat, and responds to a handler's tension or calm with visible changes in its own behaviour. In equine-assisted psychotherapy this immediacy is the point: the horse externalises a client's arousal faster and less ambiguously than a clinician's inference can, giving an in-the-moment, non-verbal biofeedback signal that anchors work on emotional regulation (#ref-selby-2013).

The horse as an affective mirror

A horse reads a handler’s autonomic arousal from posture, breathing, and movement, and its own reactivity tracks it. Move the handler’s arousal; the horse gives immediate, unfakeable feedback the therapist and client can both read.

Horse: Settled — soft eye, lowered head, licking and chewing. Resting heart rate ≈ 33 bpm.

Because the response is threshold-shaped, a client learns that small reductions in their own arousal, once past the horse’s tipping point, visibly settle the animal — a live biofeedback loop rather than a verbal report.

The second mechanism is purely sensorimotor and specific to mounted work. At the walk a horse's back delivers roughly a hundred rhythmic, three-dimensional displacements per minute to the seated rider, driving the pelvis through an oscillation close in pattern to the one produced by human walking. Hippotherapy uses this input as graded proprioceptive and vestibular stimulation to target postural control, trunk stability, and gait, particularly in cerebral palsy and other motor conditions (#ref-white-lewis-2020). The sensorimotor claim is mechanistically the most concrete the field makes, because the input can be measured and dosed in a way the relational effects cannot.

Forms and Applications

The applications sort onto the two channels. Hippotherapy and therapeutic riding exploit the gait for physical and developmental goals: gains in gross motor function among children with cerebral palsy are the field's best-supported sensorimotor outcome. Bass and colleagues, in a controlled trial of therapeutic riding for children with autism, found improvements in social functioning and sensory processing relative to a waitlist, illustrating how the mounted format is also used toward non-motor targets (#ref-bass-2009).

The ground-based relational applications reach psychiatric and psychological populations. Nurenberg and colleagues ran one of the field's few randomised trials, comparing equine-assisted psychotherapy and canine-assisted therapy against standard care in chronic psychiatric inpatients, and found reduced violent incidents in the equine condition among patients with a history of aggression (#ref-nurenberg-2015). Programmes for anxiety, depression, trauma, and at-risk youth have proliferated on the same relational logic, though the strength of their evidence varies sharply with the rigour of the design (#ref-kendall-2015).

Table 1

The four modalities of equine-assisted therapy, sorted by channel

The Evidence Base

Systematic reviews give a cautiously positive but heavily qualified signal. Selby and Smith-Osborne synthesised the controlled equine literature across physical and psychosocial outcomes and reported benefits alongside pervasive methodological weakness — small samples, weak controls, and inconsistent measures (#ref-selby-2013). Kendall and colleagues, restricting attention to psychological outcomes, reached a similar verdict: a pattern of favourable results undercut by designs too weak to license a confident causal claim (#ref-kendall-2015).

What a moderate effect survives at a real sample size

Set a point estimate and a per-arm sample size. Cohen’s U₃ and the probability of superiority translate the effect into plain terms; the confidence interval shows whether the same data can be told apart from no effect at all.

U₃ = 67.4% of treated clients exceed the average control outcome. Probability of superiority = 62.5%.
SE(d) ≈ 0.286; 95% CI = [-0.11, 1.01].

The interval includes zero: at this sample size a moderate benefit cannot be told apart from none.

The recurring problem is not the direction of the effects but their credibility. Where trials are randomised the effects tend to shrink, and where they are not, expectancy and selection are hard to exclude — a patient who chooses to sit on a horse differs from one who does not. The evidence base is therefore best read as a set of promising point estimates whose confidence intervals, at the sample sizes achieved, frequently include no effect at all.

The Efficacy Debate

The skeptical case is precise. Anestis and colleagues reviewed equine-related treatments for mental disorders and concluded that no study to date supplied the dose-response evidence or the controls a claim of efficacy requires; they judged the manipulations too poorly specified and the outcomes too vulnerable to bias to support the enthusiasm surrounding the field (#ref-anestis-2014). The critique is the equine instance of a general pattern in animal-assisted work: Marino and Lilienfeld reached the same verdict for dolphin-assisted therapy, faulting the designs for exactly the confounds — no blinding, weak comparators, undefined dose — that recur here (#ref-marino-lilienfeld-2007).

Kazdin framed the constructive response. Rather than dismiss animal-assisted intervention, he set out the methodological programme it must complete to earn an evidence base: specify the active ingredient, standardise the dose, use active control conditions, and pre-register outcomes, so that a field rich in testimonials can generate the controlled tests it has mostly lacked (#ref-kazdin-2017). The honest summary is that equine-assisted therapy has a plausible dual mechanism and a body of favourable but fragile studies, and that the burden now sits with better trials rather than with more programmes.

Worked Example

The efficacy debate becomes concrete when an effect size is carried through to what it implies at a realistic sample size. Take a moderate point estimate, Cohen's d = 0.45, near the upper end of what controlled equine-assisted trials tend to report, from a two-group study with n = 25 per arm — itself larger than many trials in the field.

Two intuitive re-expressions of the point estimate follow. Cohen's U₃, the proportion of the treated group exceeding the average untreated outcome, is the standard-normal cumulative probability Φ(d): Φ(0.45) = 0.674. About 67% of treated clients therefore end above the average control outcome, against the 50% baseline. The probability of superiority — the chance a randomly chosen treated person outscores a randomly chosen control — is Φ(d / √2) = Φ(0.318) = 0.625, about a 5-in-8 chance.

Now the sample size bites. The standard error of d is approximately √(2/n + d²/(4n)) = √(2/25 + 0.2025/100) = √0.0820 = 0.286. A 95% confidence interval is 0.45 ± 1.96 × 0.286 = [−0.11, 1.01]. The interval includes zero: the same data consistent with a moderate benefit are also consistent with none. This is the skeptics' argument in one line — not that the point estimates are unfavourable, but that at the field's typical sample sizes they cannot be told apart from no effect.

Discussion

Equine-assisted therapy is defined by a genuine asset and a matching liability. The asset is mechanistic richness: the horse is the rare therapy animal that supplies both a measurable sensorimotor input and an unusually legible affective mirror, so the field can point to plausible pathways for both its physical and its psychological applications. The liability is that this richness has not been converted into rigorous outcome evidence, and the very features that make the intervention compelling — a large, engaging animal no participant can be blinded to — are what make it hard to test.

The defensible position treats equine-assisted therapy as a mechanism-plausible adjunct with its firmest support in the sensorimotor domain, where hippotherapy's graded gait input targets motor outcomes that can be measured and dosed, and its most contested support in the relational domain, where the psychotherapeutic claims outrun the trials. Its ceiling is set less by its biology than by the cost and difficulty of the controlled studies needed to prove what that biology delivers — and by the practical constraints of working with a large animal, from the welfare of a prey species whose willing participation is a precondition to the management of zoonosis and physical-safety risk that any clinical contact with horses entails.

Current Directions

The field's current work is an attempt to answer the skeptics on their own terms. The clearest example is the move to manualisation: Arnon and colleagues developed a structured, replicable equine-assisted therapy protocol for veterans with post-traumatic stress disorder and reported preliminary reductions in symptoms, precisely the specify-the-ingredient, standardise-the-dose step that earlier critiques demanded (#ref-arnon-2020). A manual makes an intervention testable, transferable, and — for the first time in much of this literature — poolable across sites.

A parallel effort works on the concept itself. White-Lewis's concept analysis pushes the field to fix its terminology, separating hippotherapy from therapeutic riding from ground-based psychotherapy so that reviews stop averaging over interventions that share only a species (#ref-white-lewis-2020). Both moves serve the same end that Kazdin set for the whole animal-assisted enterprise: to convert a domain rich in mechanism and testimonial into one that can meet the ordinary standards of clinical evidence (#ref-kazdin-2017). Near-term progress will be judged by the quality of the controls, not the volume of the claims.

Common Misconceptions

Hippotherapy and equine-assisted psychotherapy are the same thing.

They are different interventions with different mechanisms. Hippotherapy uses the horse's gait as a sensorimotor treatment tool, usually mounted; equine-assisted psychotherapy is a relational, almost always ground-based mental-health intervention. Averaging over the two is a chief source of unpoolable evidence (#ref-white-lewis-2020).

The benefits are proven because so many programmes report success.

Volume of testimonial is not evidence of efficacy. Systematic reviews restricted to controlled designs find the effects shrink and their confidence intervals often include zero; the field lacks the dose-response and blinded comparisons a causal claim needs (#ref-anestis-2014).

The horse must be ridden for the therapy to work.

Most psychological equine work is done from the ground, not mounted. The relational and biofeedback effects depend on interacting with the horse, not on riding it; mounting is specific to the sensorimotor applications (#ref-selby-2013).

Glossary

Cohen's d.

A standardised effect size expressing a difference between two group means in pooled standard-deviation units; roughly 0.2, 0.5, and 0.8 denote small, medium, and large effects.

Confidence interval.

A range of values, computed from data, that would contain the true effect on a specified proportion of repeated samples; an interval spanning zero indicates an effect not distinguishable from none.

Cortisol.

The principal human stress hormone; a decline during calm animal contact marks down-regulated stress-axis activity.

Equine-assisted learning.

A ground-based, horse-involved intervention directed at educational and personal-development objectives rather than clinical ones.

Equine-assisted psychotherapy.

A relational mental-health intervention, almost always conducted from the ground, in which a horse's behaviour is used toward psychological goals under a credentialed clinician, usually with an equine specialist.

Equine-assisted therapy.

The MeSH-recognised family of goal-directed therapeutic interventions involving horses; the sole narrower descriptor beneath animal-assisted therapy.

Hippotherapy.

Treatment that uses the three-dimensional movement of a walking horse as graded sensorimotor input, applied by a physical, occupational, or speech therapist toward a rehabilitation goal.

Human-animal bond.

The mutually beneficial, dynamic relationship between people and animals that equine-assisted therapy deliberately recruits for therapeutic ends.

Oxytocin.

A neuropeptide implicated in social bonding, trust, and stress buffering that rises during affiliative human-animal contact.

Probability of superiority.

The chance that a randomly chosen treated individual scores better than a randomly chosen control, computed as Φ(d/√2).

Proprioception.

The sense of the body's position and movement in space; the graded input hippotherapy targets through the horse's gait.

Therapeutic riding.

Adapted instruction in horseback-riding skills for people with disabilities, pursued for functional, developmental, and recreational benefit rather than as a graded therapy technique.

Vestibular system.

The inner-ear sensory system that registers head position and acceleration; the rhythmic motion of the horse supplies it with graded input alongside proprioception in hippotherapy.

Zoonosis.

An infection transmissible between animals and humans; managing zoonotic risk is a standing safety constraint in clinical animal-assisted work.

Key Researchers

Michael D. Anestis (living). Clinical psychologist and Executive Director of the New Jersey Gun Violence Research Center at Rutgers University; lead author of the 2014 systematic review concluding that equine-related treatments for mental disorders lack empirical support. Wikipedia

Aubrey H. Fine (living). Professor emeritus at California State Polytechnic University, Pomona, and editor of the Handbook on Animal-Assisted Therapy, the reference that codifies animal- and equine-assisted intervention standards. Faculty page

Boris M. Levinson (1907-1984). Child psychologist at Yeshiva University who introduced the modern clinical concept of pet therapy, the conceptual root that the equine field inherits. Wikipedia

Scott O. Lilienfeld (1960-2020). Clinical psychologist at Emory University and a leading authority on pseudoscience in psychology; co-author of the 2014 critique of the empirical basis for equine-related treatments. Wikipedia

Yuval Neria (living). Professor of medical psychology at Columbia University and director of its PTSD Research and Treatment Program; senior author of the manualised equine-assisted therapy for veterans with post-traumatic stress disorder. ORCID

Frequently Asked Questions

What is the difference between hippotherapy and equine-assisted psychotherapy? Hippotherapy uses the horse's rhythmic gait as a sensorimotor treatment tool, applied by a physical, occupational, or speech therapist toward a rehabilitation goal, usually with the client mounted. Equine-assisted psychotherapy is a relational mental-health intervention, almost always conducted from the ground, in which the horse's behaviour supports psychological work (White-Lewis, 2020).

How can a horse act as a mirror for emotions? As a prey animal, a horse continuously monitors posture, breathing, and movement for signs of threat and responds to a handler's tension or calm with immediate, visible changes in its own behaviour. That reactivity gives an in-the-moment, non-verbal signal of a client's arousal that clinicians use to anchor work on emotional regulation (Selby & Smith-Osborne, 2013).

Is the horse ridden during equine-assisted therapy? Not usually. Most psychological equine work is done from the ground; the relational and biofeedback effects depend on interacting with the horse rather than riding it. Mounting is specific to the sensorimotor applications such as hippotherapy and therapeutic riding (Selby & Smith-Osborne, 2013).

Is equine-assisted therapy proven to work? Controlled trials report moderate benefits for some autism, psychiatric, and trauma outcomes, but systematic reviews find the evidence limited by small samples, weak controls, and undefined dose. The effects are promising but not yet established, and a skeptical literature argues they cannot be distinguished from no effect at current sample sizes (Anestis et al., 2014).

What is hippotherapy actually treating? Hippotherapy uses the roughly hundred rhythmic, three-dimensional pelvic displacements per minute that a walking horse delivers to a rider as graded proprioceptive and vestibular input, targeting postural control, trunk stability, and gait, most robustly in children with cerebral palsy (White-Lewis, 2020).

Which populations is it used with? Sensorimotor applications serve children with cerebral palsy and other motor and developmental conditions, including autism; ground-based relational applications reach people with anxiety, depression, trauma, psychiatric illness, and at-risk youth. Evidence quality varies sharply across these uses (Kendall et al., 2015).

Why do skeptics doubt the evidence? Participants cannot be blinded to a horse, control conditions differ widely across studies, dose is rarely specified, and samples are small, so favourable point estimates carry confidence intervals that often include zero. The critique targets the design, not the direction, of the findings (Anestis et al., 2014).

Is the field doing anything to strengthen its evidence? Yes. Recent work manualises the intervention so it can be replicated and pooled (for example a structured protocol for veterans with post-traumatic stress disorder) and presses the field to fix its terminology so reviews stop averaging over unlike interventions (Arnon et al., 2020).

References