The Relationship Between Child Language Proficiency and Language of Treatment on the Outcomes of Bilingual Children With Developmental Language Disorder

The Relationship Between Child Language Proficiency and Language of Treatment on the Outcomes of Bilingual Children With Developmental Language Disorder

The Relationship Between Child Language Proficiency and Language of Treatment on the Outcomes of Bilingual Children With Developmental Language Disorder

Of the 12 million children in the USA growing up bilingual, about 1 million experience Developmental Language Disorder (DLD), a disorder in language learning and use. Currently there is no guidance for speech language pathologists (SLPs) as to the language of intervention for bilingual children with DLD with differing degrees of proficiency with English or Spanish. This project will examine the relationship between relative language proficiency and the language of intervention, considering monolingual intervention in English and Spanish and bilingual intervention presented by alternating English and Spanish treatment sessions with the goal of improving language outcomes and thereby strengthening long-term academic achievement.

More than 8.5 million children in the USA speak Spanish at home with about a half million experiencing Developmental Language Disorder (DLD), a disorder in language learning and use that cannot be attributed to limited language exposure, autism, intellectual disability, hearing impairment, etc. One key challenge in serving bilingual children with DLD is the mismatch between the language(s) they speak and the availability of Speech Language Pathologists (SLPs) who can provide services in those languages. While it seems self-evident that a monolingual child should be treated in their first language, currently there is no guidance for SLPs as to the language of intervention for bilingual children. Hence, a critical question is what language(s) of treatment will best serve children with DLD with different proficiency profiles in their development of both Spanish and English. The first question is whether gains in the treated language(s) are influenced by the child's proficiency in each language (Aim 1). Cross-linguistic transfer has been documented in priming studies suggesting that underlying syntax representations are interconnected. Transfer effects may make it possible for a child to improve in both languages as a result of treatment in one language, provided that the child has adequate levels of knowledge to connect the information provided in treatment across both languages. The clearest evidence of transfer can be derived from assessing gains in the untreated language when treatment is presented monolingually (Aim 2). Our own preliminary data suggest that recast therapy can result in gains in both English and Spanish for children treated in just one language. In this study, the investigators carry out a randomized controlled trial, enrolling 120 children with DLD between the ages of 4 and 6 who score below 40% correct on the use of conditional adverbial clauses (if-then) and Complement clauses (e.g., he wonders who will be there…). Children receive one of three possible treatments (English-only, Spanish-only, bilingual) for one grammatical structure for 9 weeks, and then outcomes are re-assessed for both structures in both languages. The second grammatical structure is then treated for 9 weeks, and outcomes are assessed a third time. Comparison of different treatment approaches will inform our understanding of what is the best approach to therapy for bilingual children with a particular proficiency profile. Comparison of gains across languages and targets will allow us to determine the role of cross-linguistic transfer in language learning and to inform theoretical accounts of language representation in the developing bilingual child.

developmental language disorder, bilingual children, sentence recast, Bilingual proficiency, Language transfer

Participants will be randomly assigned to one of three conditions: monolingual recast therapy delivered in English, monolingual recast therapy delivered in Spanish, or bilingual therapy (Spanish+English). Schools are randomly assigned to first treatment target (conditional adverbial clause; complement clause). The investigators use a cross-over design such that all children eventually receive treatment for all targets. Pretesting will be completed for both targets in both languages. Children will be stratified based on bilingual proficiency (Spanish-dominant, English-dominant, Balanced) and randomly assigned to language of intervention. Children will receive 16 hours of recast therapy. After, both structures will be tested in both languages again. Then, treatment target will switch but language of intervention will remain the same, and children will receive an additional 16 hours of treatment for the second structure. At post-test, both structures will be tested in both languages.

A trained, bilingual SLP will treat the targeted structure at a rate of ~ 1 recast per minute, for 16 hours spread over 9 weeks to obtain a planned dose of 912-1008 recasts (960 +/- 5%). Following evidence on enhanced conversational recasting, the SLP will obtain the child's attention before recasting and systematically vary the lexical items in the recasts. Children receiving monolingual Spanish therapy will have the entire treatment session conducted in Spanish.

A trained, bilingual SLP will treat the targeted structure at a rate of ~ 1 recast per minute, for 16 hours spread over 9 weeks to obtain a planned dose of 912-1008 recasts (960 +/- 5%). Following evidence on enhanced conversational recasting , the SLP will obtain the child's attention before recasting and systematically vary the lexical items in the recasts. Children receiving monolingual English therapy will have the entire treatment session conducted in English.

Treatment will differ from monolingual therapy in that the child will be seen by two SLPs in keeping with one-person one-language models. This allows us to ensure that the dose in each language is controlled and supports the use of both languages evenly in therapy. Sessions will alternate between English-only therapy and Spanish-only therapy - thus the child will receive 8 hours of therapy treating the selected target in English and 8 hours treating the selected target in Spanish. A child in bilingual therapy will receive approximately 456-504 (480 +/- 5%) recasts in each language for a total of 912-1008 recasts combined.

Recast therapy is a well-established treatment for grammar in children with DLD. In this treatment, the adult repeats the child's own utterance, altering it to include the taught structure. It yields consistent large effect sizes (Hedge's g = 0.7-1.0) when focused on a single target and provided at a high dose (10-20 hrs. of therapy at a rate of ~1 recast/minute or ~600-1000 recasts total) for both morphology and syntax .

Accuracy on elicited production probes are the primary outcome measure. There are 40 probes in total. Ten for Spanish conditionals, ten English conditionals, ten Spanish complement clauses and ten English complement clauses.

Language samples serve as a means to assess functional change. Narratives are more likely than conversation to elicit complex syntactic forms while still being appropriate for this age range. Children will provide one story retell in each language before, at midpoint, and after treatment. Frog books from Mercer Meyer with story-retell scripts developed for this study will be used. The investigators will obtain the number of target utterances used in each language.

Inclusion Criteria: parent concerns and/or a history of receiving services in the public schools age-specific cutoffs for the morphosyntax subtests for their best language (English or Spanish) on the Bilingual English Spanish Assessment. The cut-off score for best language for 4-year-olds is 84, for 5-year-olds is 85, and for 6-year-olds is 81. Using the best-language approach, these scores have a sensitivity over 90% and specificity over 80% for children between 4;0 and 6;11 years of age , which is considered acceptable for studies of diagnostic accuracy. nonverbal IQ, as measured by the Kaufman Brief Intelligence Test-2, matrices subtest, will be at or above a standard score of 70. pass a hearing screening test participants must be bilingual, that is children must be producing at least simple sentences in spontaneous speech in both Spanish and English participants must be able to benefit from treatment for both conditional adverbial clauses and complement clauses, as evidenced by accuracy below 40% on 10-item elicited production probes in both languages Exclusion Criteria: 1) children with significant sensory-motor concerns or psychiatric disorders per parent report will not be enrolled.

General summary data and individual-level data will be shared for the following variables: Age, parent education level, scores on standardized tests associated with eligibility and relative language proficiency, percent correct on pre-/mid-/post-test elicited production probes, and summary variables for language sample data (e.g., number of target utterances). This information will be made available in the form of a comma delimited text file and a code book. Individuals' identities by redacting birthdates and date of examination from the available records will be disguised by replacing names with coded alphanumeric values. Materials (books, probes, training tips, training videos, etc.) will be available to SLPs and other researchers upon request. Audio files cannot be fully deidentified; therefore, data at the child level will only be available with IRB approval.