1. Paediatric Language Group
Specific Language Impairment: Therapy approaches for children with an MLU <2
Paediatric Language Group

2. Paediatric Language Group: Recap
Last year we looked at impact of dosage/intensity of treatment on outcomes: What was the optimum dosage/intensity for treatment in SLI?
Outcomes:
Distributed practice is more effective than massed practice
Difficult to define how much treatment is needed (optimal intensity)
Active ingredients not clearly enough defined in the literature - ie language therapists tend to use a mixture of dosage forms (eg. Modelling, recasting etc)
We need to clearly define what therapy techniques we are using (dosage forms) before looking at intensity

3. How did we choose our question?
General consensus among the group that we are not very good at labelling the types of language therapy we use. For example, when providing phonological therapy we are often more clear at identifying the therapy techniques/approaches we are using (eg minimal pairs, core vocab approach, multiple oppositions etc). In contrast, when providing language therapy, clinicians are more likely to use a combination of different treatment techniques to work on particular goals (eg. Modelling, recasting, use of visuals, requests for imitation etc)
Here is a graph taken after one health service in Sydney’s review of their clinician’s language therapy methods. The data was taken for all open cases and recent discharges at the end of The clinician’s were asked to identify which type of therapy approach they were using for particular cases. As you can see the most common answers were "informal language stimulation" and "targeted language". Clinician's were generally able to identify that they were doing "language therapy" but they couldn't very often be more specific (only about 20% of the time). 

4. Question for this year was:
In children with Specific Language Impairment with an MLU < 2, what therapy approaches are effective?
So our findings from last years topic on dosage, indicated that we need to more clearly define our ‘dosage forms’ or the specific therapy techniques we are using. So we started with the broad question of …..

5. For children with an MLU < 2, is Mileu Treatment an effective intervention approach?
For children with an MLU <2, is recasting an effective intervention approach?
And then we narrowed that question down to look at 2 specific treatment approaches that we believed were relevant for that population and these were Mileu treatment and recasting.
The reason we chose to look at Children with an MLU of <2 was mainly due to the fact that most group members were involved with children around that age group or developmental level
Also ensured our question was more clearly defined

6. Definitions
Mileu Treatment (MT) – focuses on improving the functional use of language in natural communicative exchanges
It incorporates:
- arranging the environment
- Mileu teaching (elicitive models, requesting imitation, prompting questions, time delay prompts)
- Functional reinforcement from the adult
Target population: MLU , able to imitate, use at least 10 words
We thought it might be useful to give some brief definitions of the 2 treatment approaches
1. Arranging the environment to provide more opportunities for the child to initiate a communicative exchange.
2. Using Milleu Prompts such as elicitive models, requesting imitation, prompting questions and time delay prompts.
3. Selecting developmentally appropriate language targets.
4. The adult responds to the child’s communicative attempt by providing access to the desired object or continuing the interaction or giving feedback in the form of an expansion.
You may have heard the term Enhanced Milleu Teaching. This therapy approach emerged at a later date and is based on the original principals of Milleu teaching but combining these with Responsive Interaction. Responsive Interaction techniques include following the child’s lead, modelling, expansions and balanced turn taking.
Basically Enhance Milleu Teaching is Milleu Teaching combined with more naturalistic techniques such as recasting. We decided to focus on investigating pure MT as there are more distinct differences between it and recasting than Enhanced mileu and recasting.

7. Definitions
Recasting- Following a child’s utterance the clinician provides an immediate expansion or correction of the child’s utterance
- the recast maintains the meaning of the utterance but corrects the grammatical or syntactical errors
-provided in a conversational context
-No child response is required
- Child needs to be intelligible i.e. so the adult can provide the correct recast
- Speech or language goal that is developmentally appropriate
Target Population: prelinguistic – preschool language
Recasting is a technique that can be applied broadly as part of an overall treatment package or it can be used in isolation. Recasting involves the child initating an utterance and the adult providing an immediate expansion of a more a sophisticated utterance. Research suggests that it is necessary to maintain the meaning of the child’s utterance. Therefore the child needs to be intelligible so that the adult can provide a semantically appropriate recast. Recasting can involve choosing specific semantic or syntactic goals to recast or it can be used as a general strategy throughout the intervention. Goals need to be developmentally appropriate and could include grammatical morphemes, complex sentence forms, vocabulary targets or individual sounds.
Recasting has been used with children with a variety of impairments such as SLI, cognitive impairment, down syndrome and asd.
The main differences between the 2 treatment approaches are that MT has more explicit prompting strategies to promote production. It’s based on the idea that practice in producing the target form is necessary to master the form.
Whereas recasting relies more on children’s spontaneous productions and does not consider production of the target as being as important, its more about the child hearing a comparison between their utterance and the adult form

8. Current Question - Literature Search
Nelson & Camarata (1996) Effects of imitative and conversational recasting treatment on the acquisition of grammar in children with specific language impairment and younger language-normal children. Journal of Speech & Hearing Research, 39, 4,
Yoder, Molfese & Gardner (2011). Initial mean length of utterance predicts the relative efficacy of two grammatical treatments in preschoolers with specific language impairment. Journal of Speech, Language, and Hearing Research, 54,  
Camarata, Yoder & Camarata (2006) Simultaneous treatment of grammatical and speech-comprehensibility deficits in children with Down Syndrome. Down Syndrome Research and Practice, 11 (1), 9-17.
Trent-Stainbrook, A., Kaiser, A. & Frey, J (2007). Older siblings use of responsive interaction strategies and effects on their younger siblings with Down Syndrome. Journal of Early Intervention, 29, 4,
When conducting the literature search we used terminology such as ‘mileu teaching’, ‘recasting’ , ‘responsive interaction’, ‘language disorders’, ‘language therapy’ , ‘children’
-Some of the difficulties we encountered in our literature search were: finding articles that used the treatment approaches, however did not fit into our criteria of having MLU<2
-Found it particularly difficult to find studies on use of recasting with children with an MLU<2 – most studies that used recasting as a technique were more advanced than this
-Some of the studies in support of recasting or EMT did not use the approach in isolation but used it as part of a broader intervention package making it difficult to determine the specific effects of the approach

9. Current Question - Literature Search
Kaiser, A., Dickinson, D., Roberts, M., Darrow, C., Freiberg, J., Hofer, K. (2011). The Effects of Two Language-Focused Preschool Curricula on Children’s Achievement through First Grade. SREE Conference Abstract Template.
Hassink, J. & Leonard, L. (2010). Within-Treatment Factors as Predictors of Outcomes Following Conversational Recasting. American Journal of Speech-Language Pathology, 19,
Yoder, P., Camarata, S. & Gardner, E. (2005). Treatment Effects on Speech Intelligibility and Length of Utterance in Children with Specific Language and Intelligibility Impairments. Journal of Early Intervention, 28 (1), 34-49
DesJardin, J. & Eisenberg, L. (2007). Maternal Contributions: Supporting Language Development in Young Children with Cochlear Implants. Ear & Hearing, 28,

10. Current Question- 3 Key articles
Yoder P, Kaiser A, Goldstein H, Alpert C, Mousetis L & Fisher R (1995). An Exploratory Comparison of Milleu Teaching and Responsive Interaction in Classroom Applications. Journal of Early Intervention, 19(3),
36 children in 6 different classrooms. Range of severities in language and cognitive ability. Aged between 2-7 years. Most children functioned at a 1-4 year old level
2 subgroups were formed by matching pairs of children based on ability. 4 developmentally appropriate language goals were chosen for each pair
1 group received Mileu and the other Responsive Interaction
MT: models, requests for imitation, mands (open questions, choice questions), time delay prompts
RI: target specific recasting, self talk, parallel talk, following the child’s lead
We have decided to discuss a few key articles that we feel have provided us with the most interesting information - you’ll see that the participants in the study didn’t always fit into the criteria of an MLU <2, however they were still able to provide us with information that was relevant to our question and caseloads
The purpose of this first study was to compare the 2 different treatments and determine were they effective in improving receptive and expressive language and determine whether the level of language ability would influence the effectiveness of different treatments.
The participants were 36 children who were attending 6 different classrooms. They ranged in severity, were aged between 2-7 years but were functioning at a 1-4 year level in terms of their language.
2 subgroups were formed by matching pairs of children based on ability. 4 developmentally appropriate language goals were chosen for each pair. The language goals differed significantly between pairs. For example, some pairs were working on single word level targets such as nouns or verbs and other pairs were working on complex syntax such as relative clauses or subordinate clauses.
1 group (or 1 child in each pair) received Mileu and the other Responsive Interaction.
The mileu treatment was provided in a child directed play activity, the adult went through a series of prompts (models, requests for imitation, choice questions, open ended question or time delay prompts) and faded these as appropriate.
The responsive interaction was a treatment package that we mentioned earlier in the slides. In this study it involved: the adult responding to the child’s communicative attempts, the adult using recasts to model the targets, use of self talk, parallel talk and following the child’s lead in play and conversation

11. Yoder P, Kaiser A, Goldstein H, Alpert C, Mousetis L & Fisher R (1995)
Yoder P, Kaiser A, Goldstein H, Alpert C, Mousetis L & Fisher R (1995). An Exploratory Comparison of Milleu Teaching and Responsive Interaction in Classroom Applications. Journal of Early Intervention, 19(3),
Teachers trained in the techniques implemented either mileu or responsive interaction to their class
Therapy given 4-5 days a week, mins, for 64 days. At least 3 activities per day conducted in free play, mealtimes, circle and small group
Results
Both mileu and responsive interaction effective at increasing child’s outcomes
For children with low expressive or receptive (age equivalent under months) milleu more effective than responsive interaction. Thought that elicited teaching more effective at teaching early targets (vocabulary learning and semantic relationships)
Higher language (age equivalent above months) responsive interaction was more effective
The teachers implemented either the MT or RI treatment to their class. Therapy was provided 4-5 days a week for mins for a total of 64 days. At least 3 activities were conducted per day in free play, mealtimes, circle and small group opportunities.
The results of the study were that the children with lower expressive and receptive language (age equivalent under months) responded better to the milleu teaching than the responsive interaction.
Whereas children with higher level language skills (age equivalent above months) responded better to the responsive interaction treatment

12. Article
2) Nelson & Camarata (1996) Effects of imitative and conversational recasting treatment on the acquisition of grammar in children with specific language impairment and younger language-normal children. Journal of Speech & Hearing Research, 39, 4,
Compared imitative treatment with conversational recasting
7 Children with SLI (ages years) matched with 7 language normal children (ages years)
6 grammatical targets selected for each child (3 absent, 3 partially absent). Included a range:
Eg. Aux, articles, possessives, relative clauses, passives, complex questions
Targets randomly allocated to a control condition (no intervention), imitation or recast condition
The purpose of this study was to compare the effectiveness of an imitation treatment vs a conversational recast treatment in children with SLI and in children with normal language development. This study wanted to look at whether 1 treatment was better than the other and look at how the 2 groups of children differed in acquiring grammatical forms when receiving identical intervention
Aquistion of language targets using recasting was better than imitiative or no treatment with children with initially absent or partly mastered targets.
- Camarata et al 1995 documented lower rates of recasts from parents in children with SLI compared to children with WNL language.
Interesting note that in their previous 1994 study which compared recasting and imitation with a group of SLI children that it was found that grammatical targets were acquired following an average of 63.6 recasts as compared to presentations in the imitation condition.

13. Received sessions 2x wk, avg 18 sessions in total.
Nelson & Camarata (1996) Effects of imitative and conversational recasting treatment on the acquisition of grammar in children with specific language impairment and younger language-normal children. Journal of Speech & Hearing Research, 39, 4,
Received sessions 2x wk, avg 18 sessions in total.
In every session, each child received:
-no intervention for 2 targets
-imitation intervention for 2 targets
eg. For the auxiliary:
Prompt: Child shown appropriate picture
Clinician Model: ‘The boy is running’
Request for imitation: ‘Say the boy is running’
-recasting intervention for 2 targets
eg. For regular past tense:
Child: ‘The baby talk’
Adult recast: ‘Yes, the baby talked on the phone’
Every child received individual therapy sessions twice a wk, receiving a total of 18 sessions each.
During each therapy session, each child received no intervention for 2 targets (these acted as the control and were monitored), the imitation intervention for 2 targets and the recasting intervention for 2 targets
The imitation intervention consisted of the clinician providing a prompt, model and request for imitation and gradually fading these cues as the child progressed. For example if targeting the auxiliary a relevant picture was shown, the clinican modelled and then did a direct request for imitation.
-The recasting intervention involved the clinician engaging the child in conversation during play with selected activities and targeting the specific grammatical targets as opportunities arose. There were no explicit prompts for production, imitation requests or verbal reinforcements

14. Nelson & Camarata (1996) Effects of imitative and conversational recasting treatment on the acquisition of grammar in children with specific language impairment and younger language-normal children. Journal of Speech & Hearing Research, 39, 4,
Results:
Found that children with SLI and normal language acquire language targets faster when conversational recasting treatment is used compared with imitative or no treatment.
Found that SLI children and their matched ‘normal language’ children were similar in the grammatical progress
Suggested that SLI children can sometimes learn grammatical structures as efficiently as WNL language children if language is tailored to specific language levels. Frequency issue?
The results of the study found that children with SLI and normal language acquire language targets faster when conversational recasting treatment is used compared with imitative or no treatment.
So both treatments were more successful than no treatment, but the recasting treatment was more effective
The study also indicated that SLI children and their matched ‘normal language’ children were similar in the grammatical progress
Although the result wasn’t statistically significant in this study, this could suggest that SLI children have the potential to learn grammatical structures as efficiently as Language normal children if the language is tailored to their specific language levels. So if provided with the right intensity of recasts, they may have the potential to learn as well as language normals

15. Article
3) Yoder, Molfese, Gardner (2011). Initial Mean Length of Utterance Predicts the Relative Efficacy of Two Grammatical Treatments in Preschoolers with Specific Language Impairment. Journal of Speech, Language, and Hearing Research, 54,  
Determine whether MLU at pretreatment could predict which 2 language treatments were more effective (MT vs Broad Target Recasting)
57 Preschoolers with SLI, avg MLU , at least 10 different words
Randomized to a MT group or a Broad Target Recast (BTR) group
MT group: 3 targets selected for each child. Child directed play activities. A series of prompts were used to elicit the target (models, questions, requests for imitation). Feedback provided
BTR group: child directed play activities. No specific targets selected
Received 30 mins therapy, 3x wk for 6 months
This study aimed to investigate whether the child’s pretreatment language ability (based on their MLU) could predict which treatment would be more effective in increasing the child’s rate of grammatical development: Mileu treatment vs a recasting treatment. Recasting was referred to as “Broad Target Recast’ in this study, as the recasts were provided to any utterance rather than pre-selected targets. The study also wanted to look at whether the treatment effects were maintained 4 months after the treatment ended.
The participants in the study were 57 preschoolers with SLI, aged months. All had an avg MLU between 1-3.5, had at least 10 different words and at least 20 recastable utterances.
The participants were randomly allocated to the MT group or the Broad Target recast group. In the MT group, three developmentally appropriate targets were selected for each child and replaced with new ones as the child progressed.
In the BTR group, treatment was also provided during child directed play. The clinician recasted any child utterance that was felt to benefit from a more developmentally appropriate grammatical structure. The grammatical structure provided in the recast had to be appropriate for the child’s current MLU stage.

16. Yoder, Molfese, Gardner (2011)
Yoder, Molfese, Gardner (2011). Initial Mean Length of Utterance Predicts the Relative Efficacy of Two Grammatical Treatments in Preschoolers with Specific Language Impairment. Journal of Speech, Language, and Hearing Research, 54,  
Results:
-Children with pre-treatment MLU <1.84 :
MT facilitated a growth in grammar faster than BTR
-Children with a pre-treatment MLU > 1.84:
No significant difference between MT and BTR
-Treatment effects were maintained 5 months post-treatment
The results of the study indicated 3 main findings: Children who started the treatment with an MLU of <1.84 showed faster grammatical development when treated with the MT approach than the BTR approach, whereas children with a pretreatment MLU >1.84 improved with both treatements but did not show a significant difference between MT or BTR
The third finding was the majority of children maintained their improvement after treatment and overall showed a moderate gain in grammatical development between the end of treatment and 5 months post treatment.

17. Other interesting findings….
Other populations may benefit from recasting techniques:
6 Children with Down Syndrome who received 2x wk recasting therapy over 6 months showed improvements in speech comprehensibility (overall intelligibility) and MLU (Camarata, S et al 2006)
Children with cochlear implants: Frequency of the mother’s recasts was positively associated with receptive language skills, MLU, number of words and different word types (Des Jardin et al 2007)
Recasting may be effectively implemented by people other than speech pathologists, although there are mixed results from these studies
Older siblings of children with Down Syndrome can be trained to use ‘responsiveness interaction strategies’ (which includes recasting techniques). Success in some participants with increasing the frequency of social commenting (Trent-Stainbrook et al, 2007)
Large number of studies have looked at effectiveness of training parents and teachers

18. Other interesting findings….
The frequency/rate of recasting is important- Children with SLI need to be exposed to a high rate of recasts in comparison to typically developing children. (Proctor, Williams, Fey, 2001)
-need to consider frequency of recasts
-need to consider giving parents specific goals re frequency
Different variables within the recasting process can influence the effectiveness of recasting (Hassink et al 2010)
-recasts are more effective if the child’s utterance contains a subject eg. Child: ‘man drinking’
Adult: ‘The man is drinking’
More effective than
Child: ‘drinking’
Recasting subject-less utterances may place additional processing on the child and be less effective
The frequency/rate of recasting is important- Children with SLI need to be exposed to a high rate of recasts in comparison to typically developing children. The frequency of recasts that parents provide naturally in conversation is not enough to warrant a change

19. Other interesting findings….
Recasts don’t always have to occur following an utterance where the child makes an error on the target.
Eg. Target: 3rd person singular
Child: That guy drinking again
Adult: He drinks his milk everyday
Recasts don’t always have to follow a child utterance. The adult can self-recast following their own utterance
Eg. Adult: It’s a cow
Child: No response
Adult: It’s a cow drinking

20. E3BP
Clinicians in language group tested out ‘techniques’ used during single activity/goal.
Goal: he/she acquisition
Clinicians in the group each recorded a session and were found to use a mix of the following techniques – recasting, sentence completion, phonemic cueing, imitation and modeling
Also trialled a session in which they solely used recasting. Found it very difficult to only use recasting especially with children who had more severe language skills or had no mastery of the target (as clinician had to self-recast the whole activity)
Found that we tended to use recasting with children with higher language levels as self-recasting only feels unnatural
What did we learn? We are using an eclectic mix of techniques – good to be more aware that recasting will be more effective with children with a MLU < 2. Clinicians in group now have a better understanding of what Milleu teaching is and how to use it in practice/appropriate candidates for the technique.
For our E3BP component this year we decided to look at our own therapy techniques and determine:
How frequently were we using recasting as a technique?
In what situations were we using it?
Was it possible to use ‘pure’ recasting as a therapy technique or was it difficult to isolate?

21. Clinical bottom line
MT and recasting are both effective therapy techniques for children with SLI and MLU<2
However…
Lower level children may benefit more from MT
MLU <1.84 and/or
Language ability lower than months
Higher level children may benefit more from recasting
MLU >1.84 and/or
Language ability above months
So our clinical bottom line from this topic was that MT and recasting are both effective therapy techniques for children with SLI who have an MLU of less than 2.
However an interesting finding was that children who have lower language skills (ie have an MLU <1.84 and language ability lower than months) may benefit more from MT
And higher level children (with an MLU >1.84 and/or language ability above months) may benefit more from recasting

22. Clinical Bottom Line
Why is MT potentially more effective for lower level children?
Production practice may be more effective than listening to models alone (Connell, 1987; Ezell & Goldstein, 1989)
Recasting may be more difficult to do with lower level children because:
They talk less frequently
The child’s intended message may not always be clear
Why is recasting potentially more effective for higher level children?
More complex syntax may be difficult to teach using MT
Children with more advanced language provide more frequent utterances for the adult to recast on
Why could this be the case? Some of the studies suggest that MT may be more effective for the younger population because it uses elicited production methods and children may learn better from saying a target rather than just listening to the language model.
There is also the suggestion that it is more difficult to implement recasts with lower level children as they talk less frequently which prevents the adult from using recasts as frequently. If the child isn’t talking much, the direct prompt in MT may be more effective
Another suggestion is that when the child is at single word level the meaning of the utterance can sometimes be unclear and to be effective, a recast has to actively encode the child’s message.
In regards to why recasting is more effective for children with more advanced language skillls- it was felt that more complex syntax may be difficult to teach using MT and that the child may not have the working memory capacity to imitate longer, more complex structures. It was felt that pairing the child’s utterance with a similar but more advanced structure, helped the child notice the new semantic or syntactic information and realize what was being taught
Recasting is also easier for higher level children as they typically provide more frequent utterances for the adult to recast on.

23. In Summary Things to consider: Child’s MLU
Choosing specific targets during recasting
Adequate rate/intensity of recasts or MT episodes
The types of recasts provided eg. Avoiding recasting too much information, avoiding recasting on utterances where the intended message is not clear
So what we’ve learnt from our clinical bottom line and our E3BP is that recasting and MT may both be effective techniques that we can apply to SLI kids on our caseload. But we need to consider the child’s MLU when selecting which treatment is appropriate. We also need to think about choosing specific targets rather than broad targets during recasting (as generally these were found to be more effective).
We need to think about ensuring the rate/intensity of recasts or MT episodes is as high as needed and monitoring this in our own practice as we still don’t know what is ideal for different populations.
We also need to consider the types of recasts – how much information is being provided and whether it’s too much for the child to process

24. Thank you for listening See you next year!!!