Pet Therapy in the Emergency Department and Ambulatory Care: A Systematic Review and Meta-Analysis

BACKGROUND: Pet therapy, or animal-assisted interventions (AAI), have demonstrated positive effects for patients, families and health care providers (HCP) in inpatient settings. However, the evidence supporting AAI in emergency or ambulatory care settings is unclear. We conducted a systematic review to evaluate the effectiveness of AAI on patient, family, and HCP experience in these settings. METHODS: We searched (from inception to May 2020) Medline, Embase, Cochrane CENTRAL, PsycINFO, and CINAHL, plus grey literature, for studies assessing AAI in emergency and ambulatory care settings on: 1) patient and family anxiety/distress or pain; and 2) HCP stress. Screening, data extraction and quality assessment were done in duplicate with conflicts adjudicated by a third party. Random-effects meta-analyses are reported as mean differences (MD) or standardized mean differences (SMDs) and 95% confidence intervals (CIs), as appropriate. RESULTS: We included 9 randomized controlled trials (RCTs; 341 patients, 146 HCP, 122 child caregivers), 4 before-after (before-after; 83 patients), and 1 mixed-method study (124 patients). There was no effect across three RCTs measuring patient-reported anxiety/distress (n=380; SMD: -0.36, 95% CI: -0.95 to 0.23; I² =81%), while two before-after studies suggested a benefit (n=80; SMD: -1.95, 95% CI: -2.99 to -0.91; I² =72%). Four RCTs found no difference in measures of observed anxiety/distress (n=166; SMD: -0.44; 95% CI: -1.01 to 0.13; I² =73%) while one before-after study reported a significant benefit (n=60; SMD: -1.64, 95% CI: -2.23 to -1.05). Three RCTs found no difference in patient-reported pain (n=202; MD: -0.90; 95% CI: -2.01 to 0.22; I² =68%). Two RCTs reported positive but non-significant effects on HCP stress. CONCLUSIONS: Limited evidence is available on the effectiveness of AAI in emergency and ambulatory care settings. Rigorous studies using global experience-oriented (or patient-identified) outcome measures are required.