Calm with horses? A systematic review of animal-assisted interventions for improving social functioning in children with autism.

No study was rated low risk of bias across all the areas the reviewers checked, and the reviewers found no trial that properly blinded the people assessing outcomes. Outcomes were mostly rated by a parent, carer or teacher who knew whether the child got the horses, so their expectations could colour the scores. Four of the eight trials compared the animal programme with a waitlist, which can make the improvement look larger than it is and cannot be blinded. The review combined results in writing rather than pooling the numbers, so it cannot tell you how big any improvement was. It was also not registered before it started, so nobody can check whether the methods changed along the way. Samples were small: nine children in the dog trial and several others under 30, so some studies may have been too small to detect a real difference. Many trials excluded children with intellectual disability, and one excluded children whose seizures were not controlled, even though epilepsy is more common in autistic people than in the general population. The results may not apply to children with more complex needs. Only one trial reported the children's IQ, and trials differed in whether they included verbal or non-verbal children.

The authors declared no conflicts of interest in relation to the research, authorship or publication of the article, and said they received no financial support for it.

Do not read this as proof that animal-assisted therapy works. The only dog-assisted reading trial found no improvement in social skills, the dolphin programme's gains matched a control with no animals in it, and none of the horse trials was free of risk of bias, most often because the person rating the child knew which group the child was in.