Reflections on Adapting an Early Language Programme to Languages Other than English While Maintaining Active Ingredients

Building Early Sentences Therapy (BEST) increases 3- to 6- year-old children's comprehension and production of simple sentence structures by promoting abstract representations of underlying predicate argument structures (PAS). It has been shown to be effective with monolingual English-speaking children. Important active ingredients include joint action routines, modelling to support mapping of meanings to PAS, and systematic variation of nouns and verbs within sentence frames. Number of exposures to a PAS and variation between sentence content are carefully controlled to ensure an intensive dosage. We describe the adaptation of BEST to languages other than English (LOTE), preserving the sentence models and variation that support these ingredients.

To adapt BEST for languages other than English, remaining faithful to its underlying theoretical principles and retaining its active ingredients. To create a methodology for adapting interventions and related assessment materials.

Two languages were adapted, Polish and Sylheti. Sentence structures from the BEST assessment and intervention were translated morpheme-by-morpheme. Principles laid out in The New Reynell Developmental Scales Multilingual Toolkit were then used to examine how PAS is conveyed (for example through word order or through case endings supporting relatively free word order), and to ensure close adherence to the original dosage type and intensity. Vocabulary equivalence and cultural appropriateness were checked. Therapy materials were created corresponding to the previous English-based ones, retaining active ingredients for effectiveness. The adapted versions of the BEST intervention were piloted with one Polish-speaking and one Sylheti-speaking child who each had language difficulties in their home language; the children were identified through teacher/parent concern, language questionnaires, and non-English adaptations of sections of a commonly used language test.

Initial assessments of the two children suggested intervention should focus on comprehension and production of PAS in simple sentences. This was confirmed through BEST assessment results, and it was then possible to use a language-adapted BEST intervention programme for each child.

To maintain the active ingredients of this intervention, it was important to ensure that suitable analogues of the target structures were identified, and underlying thematic role mapping maintained. This made it possible to evaluate whether and how the active ingredients of the intervention would apply when using adapted materials, drawing on knowledge of the target language and any available language acquisition data. For adaptation of this kind, consideration should be given to the intervention goals, ingredients and dosage, ensuring that these do not become lost in translation. Further studies are needed on applying the intervention in two or more languages and where code switching is commonplace.

Careful adaptation enables intervention in a variety of languages, especially when this is not focussed on surface phrase order and grammar.

What is already known about this subject The BEST intervention has been evaluated with English-speaking children. However, no adaptation to date has been made for languages other than English. What this study adds to the existing knowledge To our knowledge that is the first attempt to adapt an early language programme to languages other than English. The study provides principles that enable preservation of the principles and active ingredients of the programme. What are the clinical implications of this study? The study provides a methodology for adapting interventions to different languages. It explores the resource impact of doing this while indicating why it is important nevertheless to do so.