RESUMO
Previous research has proven that the balance of autistic children is poor. However, the reliability of assessing balance in this cohort has been inadequately researched. This study therefore aimed to examine if field-based static and dynamic balance tests can be reliably assessed in autistic children, to determine the number of familiarisation sessions required and whether autistic severity impacts on the reliability of these balance tests. The balance of eighteen primary school-aged autistic children was assessed three times a week over five weeks, using the flamingo balance test, a modified version of the balance error scoring system (BESS), the low beam walking test, and the heel to toe walking test. Reliability criteria included an intraclass correlation coefficient (ICC) level of ≥0.75 and a coefficient of variance (CV%) of ≤46% for the low beam walking test, the heel to toe walking test, and the BESS, and a CV% of ≤82% or the flamingo balance test. Inter-session reliability was achieved and required the least number of familiarisation sessions for the flamingo balance test, compared to the low beam walking test, which required a greater number of familiarisation sessions to achieve inter-session reliability. The heel to toe walking test and the BESS achieved inter-session reliability and familiarisation in an acceptable time frame. Due to the large CV% values reported in the current study, practitioners need to be aware that balance interventions need to achieve improvements greater than the CV% in this cohort.
