Abstract

Background:
Early intervention is recommended for pre-school children with low language. However, few robustly evaluated language interventions for young children exist. Furthermore, in many interventions the theoretical underpinnings are underspecified and the ‘active ingredients’ of the interventions not tested. This paper presents a quasi-experimental study to test the efficacy and examine the active ingredients of Building Early Sentences Therapy (BEST): an intervention based on usage-based theory designed to support young children to understand and produce two, three and four clause element sentences. BEST manipulates the input children hear to support them to harness the cognitive mechanisms hypothesized in usage-based theories to promote the development of abstract linguistic representations. One such input manipulation is the use of signing alongside verbal input signaling both content and morphology of target sentences.
Aims:
To examine whether 1) BEST is more effective than Treatment as Usual and 2) signing of content and morphology is an active ingredient of the intervention.
Methods & procedures:
A quasi-experimental study recruited children aged 3;5-to-4;5-years from thirteen schools. Schools were assigned to receive either BEST with sign, BEST without sign or Treatment as Usual (TAU). TAU group received their usual classroom provision. Across arms schools were matched with respect to classroom oral language environment and indices of deprivation. Participants were forty-eight children (twenty-eight boys) with expressive and/or receptive language abilities ≤16th centile measured using the New Reynell Developmental Language Scales (NRDLS). Outcomes, gathered by researchers blind to treatment arm, were NRDLS production and comprehension standard scores and measures of production of targeted sentence structures.
Outcomes & results:
Primary outcomes indicate that BEST with sign was significantly more effective than TAU with respect to NRDLS production standard score but not comprehension. The advantage for production was maintained at follow-up. BEST without sign was significantly more effective than TAU on measures of targeted vocabulary, sentence structure and morphology.