Evidence from systematic reviews
Recent systematic reviews of the literature report evidence of the effectiveness of speech and language therapy interventions for expressive language outcomes for children with SLI that cannot be accounted for by low IQ, behaviour or emotional problems or hearing or neurological impairments.3,51 However, the picture for receptive language outcomes is more problematic owing to a dearth of evidence, particularly from randomized controlled trials (RCTs), the effects of early remission,4 lower incidence of RELI relative to specific expressive language delay, and variability in the criteria for eligibility for recruitment.
Law et al. in an early review51 identified only five studies with receptive language outcomes that met the eligibility criteria of controlled studies of effects of intervention upon children in the age range 0 to 7 years with ‘primary’ speech and language delay (akin to SLI but not based upon formal psychometric discrepancy criteria). Four of these studies involved children aged 36 months or younger. This raises issues about the reliability of the test scores because measures obtained from preschool children are particularly susceptible to the influence of factors associated with development, such as short attention span and distractability, levels of activity, and problems in engaging with an unfamiliar test administrator.52 In addition, few of the participants in these studies had RELI, with receptive language outcomes reported for children receiving intervention for specific expressive language delay. It is thus unclear whether these interventions would be beneficial for children with RELI.
In Law et al.’s more recent review,3 only two studies met inclusion criteria for interventions targeted on receptive language (here, receptive syntax) with ‘no treatment’ control groups, both reporting non-significant effects. In the first of these studies Glogowska et al.53 reported a non-significant standardized effect size (SES) of 0.19 (95% CI −0.12 to 0.51) from a sample of 155 preschool children (71 receiving treatment and 84 controls). In the second study, Law et al.54 reported a non-significant SES of −0.45 (95% CI −1.18 to 0.28) from a sample of 38 preschool children (28 treatment and 10 controls).
Evidence from recent randomized control trials
Evidence from four recent large-scale RCTs not thus far included in published systematic reviews report interventions for children with RELI. Three of these studies investigated interventions based upon underlying auditory processing deficits. The fourth was based upon existing models of language therapy in the UK.
Cohen et al.55 reported the findings from a multicentre, intention-to-treat RCT performed in Scotland with blind assessment of outcomes to determine the effectiveness of the Fast ForWord-Language program.56 This is a computer-based intervention that utilises auditory processing theory39 and uses games with signal-processed modified speech designed to compensate for underlying auditory temporal processing difficulties. The participants (n=77) were aged between 6 and 10 years, and were monolingual English speakers with a diagnosis of RELI. Average scores for both receptive and expressive language on the Clinical Evaluation of Language Fundamentals (CELF-3UK),57 a standardized language test, were −2SDs below the mean. The children were randomized into one of three groups: a group receiving Fast ForWord-Language, a comparison group receiving ongoing language therapy, and a second comparison group who played educational computer games with unmodified speech. Outcomes were measured at 9 weeks’ post-baseline assessment and at 6 months’ follow-up by qualified SLTs not otherwise involved in the project who were blind to the children’s research group. The results revealed no significant additional benefit from playing the Fast ForWord-Language games 5 days a week for 6 weeks under parental supervision for 90 minutes each day relative to the first control group (SES −0.04 [95% CI −0.59 to 0.52] for receptive language) nor relative to the computer-games control group, who played commercially available educational computer games without modified speech on the same schedule. This trial did not support auditory processing deficits as a general explanation of severe RELI, although this was a particularly impaired cohort of children.
A recent large-scale RCT performed in the USA58 also investigated the effectiveness of Fast ForWord-Language. Participants included children with RELI but their progress could not be distinguished among children with specific ELI.
Bishop et al.59 addressed issues relating to both auditory temporal processing and to limited general processing capacity explanations of RELI in an RCT involving 36 participants aged 8 to 13 years. The children had scores of less than −1SD on standard measures of language. Participants with RELI failed to benefit from a computer training program for comprehension of grammatical constructions to help sentence comprehension (SES 0.04 [95% CI −0.2 to 0.28]). The findings once again fail to support auditory processing deficits as a general explanation of RELI within the range of their study. They would, however, be compatible with a general limited processing capacity explanation and suggest that a more individualized, contextualized approach may be preferable for children with RELI, in contrast to the computer-based rote-learning approach used on the study.
Boyle et al.60,61 investigated the effectiveness of current language-therapy practices based upon meta-analyses of published studies.51,62 The participants in their RCT were 161 children aged 6 to 12 years who had persistent primary receptive and/or expressive language impairment with no reported marked hearing loss and no moderate/severe articulation/phonology/dysfluency problems or who otherwise required individual SLT work. Eighty-six of the children had RELI (defined using a threshold criterion of CELF-3UK Receptive Language60 standard score ≤81 and non-verbal IQ scores of >75) and 75 had specific expressive impairment. They were randomized to one of five conditions, which were as follows: (1) Individual, direct project therapy: SLT working individually with a child (n=34, 20 with RELI); (2) Group direct, project therapy: SLT working with a small group of children (n=31, 17 with RELI); (3) Individual, indirect project therapy: a trained SLT assistant working individually with a child (n=33, 17 with RELI); (4) Group, indirect project therapy: a trained SLT assistant working with a small group of children (n=32, 18 with RELI); and (5) Control group (who received existing community-based services; n=31, 14 with RELI).
Project therapy was delivered three times per week for 15 weeks, in 30- to 40-minute sessions, and those in the comparison group received their ongoing therapy regime. The therapy focused on comprehension monitoring, vocabulary development, grammar, narrative, and developing language learning strategies. All post-baseline measures were blind-assessed by qualified SLTs not otherwise involved with the project.
There was no significant difference between the four modes of project therapy but children with specific expressive impairment made greater gains in both receptive and expressive language than those with RELI (all p values <0.025). Further, although the children receiving project therapy made significant overall gains in expressive language (p=0.031), there was only a modest and non-significant intervention effect for receptive language scores relative to the comparison group for the subgroup of children with RELI (SES 0.25 [95% CI −0.32 to 0.82]). However, the impact of the small numbers involved on the statistical power of this comparison (14 in the comparison group and 72 receiving project therapy) should be noted.
Recent phase I and small-scale trials including children with established RELI also suggest vocabulary development as a promising intervention. Direct teaching of vocabulary was effective with four children aged 10 to 11 years (Easton et al.63) and two children aged 8 to 9 years using criterion-referenced measures,64 and as effective as narrative intervention in developing language skills with a cohort of 54 secondary school children with RELI.65‘Traditional’ therapy including vocabulary teaching was as effective at encouraging eight children over 8 years old with severe RELI to use a mental visualization strategy to aid their comprehension of oral narratives.66 Interestingly, mental imagery training itself produced a significant improvement in the responses of children with RELI to literal questions about a short narrative.66 Furthermore, a small-scale RCT found that developing semantic definitions of verbs was as effective as syntactic–semantic shape coding on criterion-referenced measures of verb argument structure for 27 children aged 11 to 16 years with severe RELI attending a specialist residential school.67
In the case of young preschool children, Camarata et al.68 found that a treatment group of 21 children with an mean age of 31 months with RELI made significantly greater gains in receptive language in response to an intervention focused on expressive grammar than a randomly allocated comparison group of six children (mean age 37.6mo; p<0.05, SES 1.07). The intervention consisted of twice-weekly individual sessions of an hour for 12 weeks using imitation, modelling, and conversational re-casting approaches targeted on improving production of grammar. Further investigation of such transfer effects with young children would be of interest, although it should be noted that the eligibility criteria of 1SD below expected levels on standardized measures of both expressive and receptive language would have resulted in the recruitment of some children with less severe levels of impairment. It is also unclear to what extent these problems are likely to persist over time.
Mapping practice onto theory
The extent to which professional practice with children with RELI maps on to underlying theory has been investigated in a recent survey of qualified SLTs in the UK, focusing on practice with children aged 5 to 11 years.69 The findings revealed that children with RELI are seen as a priority and receive extensive services that reflect diverse practice. This includes interventions targeted on specific deficits or based upon published programmes/frameworks for practice, behavioural approaches to teaching vocabulary, and sentence comprehension. Meta-cognitive activities (e.g. training to think about communication) were widely used, in particular with older children: nearly 80% of all activities reported with those aged 11 years compared with some 20% of reported activities for those aged 5 years. Underlying theory did not appear to be regarded as important for informing intervention, and SLTs placed more emphasis on the presenting problems associated with the child’s deficit. This begs the question of the use of current theory for informing interventions and of the effectiveness of disseminating research findings to practitioners. However, the low number of respondents (n=56) should be noted.