This poorly reported review concluded that studies were encouraging and indicated therapy or assistance dogs could promote positive behaviours in children diagnosed with autism spectrum disorders, but called for further research. The results cannot be considered reliable given the apparent variability in the small evidence base.

To provide a critical review of the literature assessing the effects of therapy and assistance dogs on children with autism spectrum disorders and their families.

PubMed, ERIC, Science Direct, Google Scholar, reference lists and review articles were searched to identify eligible peer-reviewed articles published in English. Search dates and terms were not reported.

Experimental studies, case studies and semi-structured interview studies were eligible for inclusion. Qualitative (anecdotal) studies were excluded. Participants were children aged between three and 15 years, diagnosed with a pervasive developmental disorder (which includes autism spectrum disorders) according to the Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV criteria. Interventions that involved therapy dogs (brief supervised interactions between dog, child and therapist) or assistance dogs (dog living with family) were included. Full definitions provided in the paper.

Of the included studies, two reported on placement of assistance dogs and four reported on the impact of sessions with a therapy dog (or other animals). Study country was not reported. Five of the studies focused on children diagnosed with autism spectrum disorders. Mean ages ranged from three years to 15 years. Settings included the home, school or an out-patient therapy centre. Where gender was reported, participants were more likely to be male. Assistance dogs were in place within the home for an unspecified period of time. Sessions with the therapy dogs lasted between 15 and 45 minutes, the number of sessions was not reported. Observed outcomes included levels of social interaction/behaviours and isolation, language use, negative or positive behaviours, perceived safety and salivary cortisol.

It was unclear how many reviewers selected the studies for inclusion.

Studies were not formally assessed for methodological quality.

Key study characteristics, measured outcomes and main results were extracted for each study and tabulated. It was unclear how many reviewers extracted the study data.

A narrative synthesis appeared to have been conducted with data presented as text and tables. Studies were grouped according to the nature of the intervention: assistance or therapy dog.

Six studies were included in the review (97 participants). One longitudinal study used ethological observations and semi-structured interviews with parents, two studies used pre- versus post-test designs, two studies used cross-over designs and one was a single case study. One cross-over study included two control arms (stuffed dog, ball) in comparison with a therapy dog, the second used standard occupational therapy as the control. Where cross-over designs were used it was not clear if participants were randomised for order of treatment.

Assistance dogs (two studies, 52 participants): Both studies reported positive outcomes including improved safety perceptions, increased social acknowledgments and decreased problematic behaviours. One study that reported cortisol levels found cortisol awakening response decreased when the assistance dog was present in the family, and increased when the dog was removed.

Therapy dogs (four studies, 45 participants): All interventions involved the presence of a therapist and in all four studies interaction with the dogs was observed to dampen social isolation and withdrawal. Three studies reported that therapy dogs promoted verbal and non-verbal behaviours directed towards both dog and therapist.

The studies suggested that interaction with therapy or assistance dogs promoted positive verbal and non-verbal behaviours in children diagnosed with autism spectrum disorders, but further well-designed research was required.

This review addressed a broad topic with partially defined inclusion criteria. The limited number of databases searched (which omitted resources that included behavioural science and mental health), search restrictions to studies in English, and the inclusion of only published articles may have resulted omission of relevant material. The review processes were poorly described which made it difficult to rule out reviewer error or bias.

The included studies were not assessed for bias and insufficient study details were provided. The narrative synthesis was sparse, did not report if there were any statistically significant findings in the primary papers, and failed to highlight studies with more reliable designs.

The authors' cautious conclusions are based on a small and variable evidence base drawn from a poorly reported review, and should be considered as unreliable apart from the recommendation for further research.

Practice: The authors did not state any recommendations for practice

Research: The authors stated that there was a need for basic research utilising larger sample sizes to assess the effectiveness of animal-assisted interventions using randomised controlled trial designs. In addition, studies were needed to examine whether the effects of contact with dogs are maintained without constant exposure.

Italian Ministry of Health; Istituto Superiore di Sanita; Azienda per i Servizi Sanitari n. 6 Friuli Occidentale, collaborative project.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.