There are few areas where the relationship between humans and dogs raises as many legitimate questions as therapy and animal-assisted intervention.

Testimonials abound, anecdotes circulate freely, but what about the actual scientific evidence?

Defining what we're talking about: therapy, mediation, activity

Before discussing the studies, a clarification of terms is necessary, as confusion between terms is frequent and often distorts debates.

The International Association of Human-Animal Interaction Organizations (IAHAIO) distinguishes three levels of intervention.

1. Animal-assisted activities (AAA) are informal, without formalized therapeutic objectives.
2. Educational-assisted interventions fall within a school or educational framework.
3. Animal-assisted therapy (AAT), also known as therapeutic animal-assisted intervention, is conducted by a healthcare or social care professional, with defined objectives, outcome evaluation, and a structured protocol.

In France, the term animal-assisted intervention is now preferred because it encompasses a broader framework than simple zootherapy, a historical term still used.

In terms of institutional recognition, animal-assisted intervention appears in several works and recommendations as a possible non-pharmacological intervention in certain contexts, particularly in the management of behavioral disorders.

However, it is not subject to a unified national framework or homogeneous validation across all indications and populations.

Dogs, the animal of choice

Among all animals used in these interventions, the dog holds a dominant position.

A systematic review published in Animals (Basel) in 2021 on intervention protocols with dogs confirms that the dog is the most documented species in the scientific literature on AAT.

A meta-analysis of 43 articles published in 30 journals, including 16 with impact factors, reveals that the dog is mentioned as the primary mediator in approximately 72% of the studies reviewed.

This is no coincidence.

The co-evolution between humans and dogs over several millennia has produced a species particularly adapted to interspecies social interactions, capable of reading human emotional signals and responding to them coherently.

What the studies say: biological mechanisms

One of the most important advances in recent research concerns the physiological mechanisms underlying the observed benefits. A systematic review published in Psychoneuroendocrinology in 2022, covering 129 studies selected from 13,072 identified, provides solid evidence.

The main observations measured during human-dog interactions are:

• an increase in heart rate variability,
• an elevation of oxytocin levels,
• a decrease in cortisol.

These indicators are consistent with an activation of the parasympathetic nervous system, a stimulation of the oxytocinergic system, and a reduction in the activity of the hypothalamic-pituitary-adrenal axis.

Oxytocin plays a central role here.

This neuropeptide, produced by the hypothalamus and released by the posterior pituitary, is well known for its involvement in social bonding, trust, and anti-stress effects.

A review published in Frontiers in Psychology (Beetz et al., 2012) suggests that the stimulation of the oxytocinergic system constitutes a credible mechanistic hypothesis to account for a large part of the psychophysiological effects documented during human-animal interactions.

The effects are more pronounced with a familiar dog than with an unfamiliar dog, and they tend to intensify with repeated interactions over time.

A 2025 meta-analysis published in Social Science & Medicine, focusing specifically on children and adolescents, confirms that dog interventions lasting longer than 15 minutes produce a significant reduction in salivary cortisol (effect size: 0.65; p = 0.038). Shorter interventions do not show a statistically significant effect. This result has direct implications for the design of intervention protocols.

Populations and results: a rigorous assessment

Mental health

A systematic review from 2024, published in JMIRx Med (Pandey et al.), covering 16 moderate to high-quality studies published between 2015 and 2023, documents significant improvements in several areas:

• reduction of psychiatric symptoms,
• improvement of social functioning,
• improvement of quality of life,
• improvement of cognitive functions,
• decrease in stress levels.

The populations concerned include people with depressive disorders, schizophrenia, and post-traumatic stress disorder (PTSD).

In military personnel and veterans with PTSD, the addition of a psychiatric assistance dog to usual care was associated with a decrease in PTSD symptoms, anxiety, and depression in a controlled study published in 2024.

These results remain preliminary, and larger randomized trials are needed to confirm them.

Anxiety

A 2026 systematic review, registered on PROSPERO (CRD42024494109), covering 31 studies published between 2013 and 2023 and including PubMed among the databases queried, concludes that the majority of interventions, primarily with dogs, produce significant reductions in anxiety after intervention.

Randomized controlled trials consistently show superior results to control conditions. Identified beneficiary populations include hospitalized patients, children in medical settings, people with PTSD, and acute care patients.

Dementia

A meta-analysis published in Psychiatry Research (Chen et al., 2022) covering 11 randomized controlled trials documents positive effects of AAT on behavioral and psychological symptoms of dementia, including depression and agitation.

However, results on global quality of life, cognitive function, and activities of daily living remain heterogeneous across studies, and the authors highlight recurrent methodological limitations, including small sample sizes and variable intervention durations.

Elderly people

A meta-analysis published in Worldviews on Evidence-Based Nursing (2020) covering 47 selected studies concludes that AAT has positive effects in physiological, psychosocial, cognitive, and behavioral domains in elderly people. The dog is the animal used in 57.4% of the included studies.

Autism

A 2024 systematic review covering 19 studies suggests that dog-assisted therapy has a positive impact on communication and social interaction in 84% of the included studies concerning children and adolescents diagnosed with autism spectrum disorder (ASD).

The dog is described as a facilitator of communication, capturing the patient's attention and serving as a positive reinforcer.

However, the authors emphasize the still heterogeneous methodological quality of this literature and recommend more rigorous randomized studies.

The limitations that science itself recognizes

Reading the scientific literature on AAT honestly requires mentioning its structural weaknesses.

A review of randomized controlled trials published in Complementary Therapies in Medicine (Kamioka et al., 2014) identifies 11 RCTs over the period 1990-2012, and notes that their methodological quality is generally low. Meta-analysis could not be performed due to the heterogeneity of the studies.

These limitations are widely recognized in recent literature: small sample sizes, lack of standardization of protocols, difficulty in performing a double-blind condition, diversity of populations and objectives, and lack of terminological harmonization until recently.

These criticisms do not call into question the clinical interest of AAT, but they justify that statements remain cautious and contextualized.

AAT is a complementary intervention, not a first-line treatment.

The well-being of the dog: a crucial ethical question

An aspect often overlooked in discussions about AAT is the potential cost to the dog itself.

The literature on the well-being of mediating dogs remains limited and sometimes contradictory: some studies identify signs of vigilance or fatigue in dogs engaged in intensive sessions, while others find no significant negative effect.

This developing body of knowledge nevertheless justifies a regular evaluation of the behavioral and emotional state of the mediating dog, appropriate recovery times between sessions, and rigorous selection of dogs beforehand.

A dog showing signs of anxiety or insufficiently socialized to institutional contexts cannot produce the desired therapeutic effect, and can even disrupt a session.

The human-animal pair is the basic unit of work, and its balance directly determines the quality of the intervention.

Professional framework in France

In France, there is no state diploma specific to animal-assisted intervention.

Training is based on private, associative, or university courses, with heterogeneous levels and content. The lack of a unified national framework is an active subject of debate in the sector.

From a regulatory perspective, the Public Health Code prohibits the introduction of domestic animals into hospital premises, with the exception of guide dogs for the visually impaired.

However, medical-social establishments have greater leeway: each establishment director can authorize the presence of animals within the framework of supervised activities, provided that this authorization is specified in the internal regulations.

The establishment of a solid handler-dog team is a central requirement.

A family dog, however affectionate, is not automatically fit to work in an institutional setting. Aptitude criteria evaluated include absence of aggression, behavioral stability in unforeseen situations, tolerance to medical odors, ability to remain calm for extended periods, and absence of separation anxiety.

What this tells us about dogs

This area of research reveals something fundamental about the canine species: its ability to function as a social interface between a human and a difficult context. The dog is not passive in these interactions.

It reads, it reacts, it adjusts.

It is not a therapeutic tool in the mechanical sense of the term, but a living actor whose own balance directly influences the quality of the interaction.

This is precisely what makes dog-assisted intervention demanding to implement correctly, and what distinguishes a serious approach from an informal animal presence.

Science, despite its current methodological limitations, confirms that this distinction has measurable consequences on results.

In conclusion

The scientific evidence supporting dog-assisted therapy and intervention is real, although still improvable methodologically. Biological mechanisms have been identified: cortisol reduction, oxytocin increase, parasympathetic activation.

The populations that benefit most from well-conducted interventions are documented: elderly people, psychiatric patients, children with ASD, traumatized individuals. The conditions for effectiveness are also known: structured protocol, trained professional, evaluated and monitored dog, minimum intervention duration.

Animal-assisted intervention is not a vague form of alternative medicine. It is a serious support practice, whose credibility relies on the rigor of those who practice it.

________________________________________________________________________________________________

Source:

• International Association of Human-Animal Interaction Organizations, AAA / AAE / AAT definitions
• Systematic review, Animals (Basel), 2021, on intervention protocols with dogs in AAT
• Systematic review, Psychoneuroendocrinology, 2022, on the physiological mechanisms of human-animal interactions (129 studies)
• Andrea Beetz et al., Frontiers in Psychology, 2012, on the role of oxytocin in human-animal interactions
• Meta-analysis, Social Science & Medicine, 2025, on salivary cortisol in children and adolescents
• Pandey A. et al., JMIRx Med, 2024, systematic review on mental health (depression, schizophrenia, PTSD)
• Systematic review PROSPERO CRD42024494109, 2026, on anxiety and dog-assisted interventions
• Chen Y. et al., Psychiatry Research, 2022, meta-analysis on dementia (11 RCTs)
• Meta-analysis, Worldviews on Evidence-Based Nursing, 2020, on elderly people (47 studies)
• Systematic review, ScienceDirect, 2017, on the impact of assistance dogs in children with ASD (29 studies, 1,121 participants)
• Kamioka H. et al., Complementary Therapies in Medicine, 2014, critical review of randomized trials in AAT
• Public Health Code (France), rules for animal access to hospitals
• HANDI'CHIENS, institutional data on the training and career paths of assistance dogs