Introduction

Specific Language Impairment (SLI) is diagnosed in children who evidence significant developmental difficulty which manifests primarily in the area of language. The term "language" is used here to refer to difficulty in aspects such as vocabulary and grammar, and is differentiated from "speech" impairments such as articulation impairments or stuttering. Language is an important part of almost all daily activities, social and academic. As a result, language impairments have a significant impact on children's success and well-being. This article discusses the main types of interventions available for the remediation of language impairment in childhood and their efficacy. Language impairment in childhood can occur as the primary diagnostic factor or as part of a more general developmental disorder (specific versus non-specific language impairment). In some respects, the intervention approach may vary depending on the children's diagnostic category in order to take into account differences in developmental profiles and learning styles. To a large extent, however, interventions are more similar than different across diagnostic categories, being tailored to fit the individual child's language level, cognitive ability, communicative needs and interests. This review will, however, focus on studies on children whose primary deficit is in the area of language.

Research Questions:

1) What methods are used to remediate language impairment and what are they based on?
2) Is there evidence to support the efficacy of language intervention methods?
3) Do the same methods work for children speaking different languages and for bilingual children?

1) What methods are used to remediate language impairment and what are they based on?

Once a child has been identified for intervention, an individualized therapy plan is typically formulated with specific therapy objectives based on an in-depth assessment of the child's abilities in comprehension and production across various domains of language such as vocabulary, grammar, and language use. The areas of language that are most significantly affected by language impairment vary across individual children and also within the same child over time (Conti-Ramsden & Botting, 1999). Therefore, therapy goals are tailored for children individually and are reassessed periodically. Multiple therapy targets are prioritized based on a number of factors see e.g., Paul, 2007). Thus, the order in which language structures are addressed usually follows their normal order of acquisition to ensure that the child possesses the necessary prerequisites for the new skill. Another important consideration is the functionality of the target for the child given the child's typical activities. This not only increases the child's motivation but also promotes immediate transfer of the skill to daily activities. The child's future needs are also considered and may at times be deemed more important than adherence to the developmental sequence. For example, it has been advocated that intervention for preschool children should not only focus on preschool-level conversational skills, but should include more advanced language skills aiming to prepare the children for classroom interactions and the acquisition of literacy (Fey, Catts & Larrivee, 1995). In general, the goals selected will vary depending on the age and developmental level of the child, with younger children typically addressing basic aspects of language, and school-age children and adolescents addressing, in addition, more complex aspects which are often tied to school-related needs. Thus, older children often work on aspects such as understanding their textbooks, writing a well-organized term-paper, and understanding complex uses of language such as multiple-meaning words and figurative languages such as idioms, similes, and metaphors (Nippold, 2000).

Available methods of language intervention vary in their focus and theoretical orientation (see review of various methods in McCauley & Fey, 2006; Paul, 2007). Some methods focus on the explicit teaching of rule-based aspects of language, whereas other methods emphasize the meaning of language without an explicit focus on formal aspects, under the assumption that the underlying rules will emerge with meaningful use. Intervention approaches vary as well in the breadth of their focus. Thus, approaches may target language skills broadly or may be designed to target specific sets of skills with the assumption that these are directly linked to success in language learning, for example, aspects of auditory processing (Tallal, Miller, Bedi, Byma, Wang, Nagarajan, et al., 1998). Strategies have also been developed to incorporate components directed specifically at memory and attention and the processing of linguistic information (Ellis Weismer, 2000; Gillam, Hoffman, Marler & Wynn-Dancy, 2002).

One way to categorize language intervention approaches is according to the extent to which the teaching activities are under the clinician's control, describing approaches as falling on a continuum from clinician-controlled to child-centered (Fey, 1986). The former type of approach allows the clinician to impose a focus on specific pre-selected therapy targets using activities that permit intensive training of these targets. In contrast, child-centered approaches use more naturalistic activities in which the clinician incorporates modeling and reinforcement of therapy targets within contexts that are meaningful to the child. However, at the same time, there is less opportunity for the therapist to control the focus of the session. At the midpoint of the continuum are hybrid, or semi-structured approaches which aim to include the benefits of approaches at each end.

2) Is there evidence to support the efficacy of language intervention methods?

To decide whether a given intervention method is appropriate and efficacious, one must evaluate its theoretical support as well as research evidence supporting its efficacy. The efficacy of clinical intervention is demonstrated by showing that progress is made which is directly attributable to the intervention. This can be addressed by testing groups of children, randomly assigned to treatment and no-treatment groups or to groups receiving two different types of intervention (in such studies, children in the no-treatment groups are generally offered treatment once their no-treatment period is over). Another method tracks the performance of individual children over time, beginning with a baseline of pre-treatment performance and subsequently demonstrating that increased performance coincides with the onset of treatment (see review in Cleave, 2001). Yet another approach rests on the demonstration that progress over a period of time exceeds that which would be expected by development alone by comparison with normative data on typical development. It must be kept in mind, however, that children with language impairment progress more slowly than typically developing children. Therefore, even if their progress in therapy does not exceed that seen in typically developing children, it may constitute a significant improvement over no treatment (Leonard, 1998). In the demonstration of accuracy, it is important as well to document not only the immediate outcome of the intervention, but also how well the gains transfer to other situations (referred to as generalization), and how well they are maintained over time. Indeed, generalization has been a notorious area of difficulty in language intervention, resulting in an increased focus on strategies that promote meaningful learning and long-term retention.

A number of published intervention efficacy studies support the efficacy of various treatment approaches, providing clear evidence that language intervention can make a significant difference for children's language development, including approaches that use highly clinician-directed methods and ones that use a more naturalistic or hybrid method (Camarata, Nelson & Camarata, 1994; Connell, 1987; Connell & Stone, 1992; Culatta & Horn, 1982; Ellis Weismer, Murray & Branch, 1989; Fey et al., 1993; Fey, Cleave & Long, 1997; Kouri, 2005; Leonard, Camarata, Brown & Camarata, 2004; Nelson, Camarata, Welsh, Butkowski & Camarata, 1996; Schwartz, Chapman, Terrell, Prelock & Alpert, 1992). These studies have used controlled group designs as well as single-subject designs and have included interventions addressing various aspects of language, including lexical and grammatical goals as well as pragmatic aspects of language. Many of the available studies have focused on preschool-age children, however, studies have targeted school-age children as well (Culatta & Horn, 1982; Parsons, Law & Gascoigne, 2005; Swanson, Fey, Mills & Hood, 2005).

The available research, therefore, indicates that language intervention works. However, there is little clear-cut evidence supporting the superiority of one type of approach over another or providing guidelines as to which approach will work best for individual children (e.g. Kouri, 2005; Law, Garrett & Nye, 2004; Leonard, 1993). Several studies have concluded, however, that interventions that emphasize learning in meaningful contexts do result in better generalization (Camarata et al., 1994; Kouri, 2005; Law, 1997). Meta-analyses of intervention studies have been conducted, which combine results across studies in order to offer a synthesis of available findings (e.g. Law et al., 2004). These studies have confirmed the efficacy of intervention for various aspects of language, but also raise important questions regarding research methods and the correspondence between intervention conducted for research and in clinics. Most available intervention studies follow children over relatively short periods of time. However, clinical experience as well as research indicates that most children with language impairment will require long term intervention (Fujiki & Brinton, 2005; Leonard, 1998). Interventions have been proposed that promise very rapid gains in language skills and some of these have gained considerable popularity. However, such intervention approaches in general remain controversial theoretically as well as empirically, when treatment gains are considered in tightly controlled research studies (e.g.; Cohen et al., 2005; Gillam, Crofford, Gale, & Hoffman, 2001; Tallal et al., 1998).

Parents often play an important role in their children's intervention. Parent involvement may be indirect, involving general support or homework assignments to support therapy activities. In other cases, parents may be trained to play a more direct role in implementing therapy activities. Studies have shown that trained parents can be effective agents of intervention (Fey et al., 1993; Proctor-Williams, Fey & Loeb, 2001; Tannock & Girolametto, 1992). However, the results of parent intervention are found to be less consistent than those achieved by a clinician (Fey et al. 1993). Recent results also support the viability of the use of trained child care workers as agents of intervention (Girolametto, Weitzman & Greenberg, 2003; 2004). These results suggest that parent-based intervention should not be seen as a replacement for clinician-based intervention. As well, it is important to recognize that not all parents are capable of assuming this role due to various factors and should not be given the primary responsibility of intervention.

3) Do the same methods work for children speaking different languages and for bilingual children?

The great majority of published intervention efficacy studies have focused on monolingual speakers of English. To what extent are the findings applicable to children speaking other languages? Clinical experience supports the efficacy of language interventions conducted in other languages, suggesting that many of the factors that facilitate language are the same across languages. However, it is also important to keep in mind that important cultural differences exist in child rearing practices, including how parents and other adults speak to children and the ways in which children are expected to respond (Brice Heath, 1983; Ochs, 1988). Therefore, further research may uncover cross-linguistic differences in optimal methods for language facilitation (van Kleeck, 1994). As a general guideline, methods using intervention styles that feel awkward to native speakers of a language are not likely to be the ones most conducive to results. Very few published efficacy studies have focused on bilingual children. A central question that arises for this group is frequently which language to use in intervention. The limited evidence available for bilingual children with language impairment suggests that intervention conducted bilingually produces somewhat better results than monolingual intervention (Elin Thordardottir, Ellis Weismer & Smith, 1997). Thus, the evidence speaks against the elimination of one of the languages, showing that it is not a helpful strategy and may in fact be counterproductive (Elin Thordardottir, 2006). Indeed, many authors have emphasized the benefit of intervention methods which strengthen both languages of bilingual children (Elin Thordardottir, 2006; Elin Thordardottir et al., 1997; Kohnert, Yim, Nett, Kan & Duran, 2005; Restrepo, 2005).

Directions for Further Research

To summarize, a number of well controlled research studies have provided evidence that language intervention with children using several different methods is efficacious. However, many aspects of intervention need to be investigated in greater detail with respect to their effect on efficacy. First, in spite of the considerable research base available, much more research is needed to further confirm the efficacy of language interventions and to better establish which aspects of these methods are the key contributors to treatment gains. Also, although there is evidence that interventions of greater duration in general provide better results (Law et al., 2004), little detailed information exists on the optimal intensity of intervention (length of sessions, number and spacing of sessions). Other questions include whether there are identifiable periods when children are more or less receptive to intervention, whether certain characteristics of children in terms of their language impairment or other ability factors make them ideal candidates for particular types of intervention, and what methods are most efficacious for bilingual children.