Communication Intervention in Real-Life Settings: Outcomes and Challenges

Background :  The Picture Exchange Communication system (PECS) is a widely used programme to enhance communication skills in non-verbal children with autism.  However, despite being used in over half of all specialist autism schools in the UK (National Autistic Society statistics, 2005), the majority of teachers using this system are neither specially trained, or monitored in their use of PECS. Thus, the quality of intervention offered to pupils is often poor. Moreover, although children who remain non-verbal through elementary school tend to have the lowest levels of intellectual and language ability, little is known about the specific characteristics of those children who do or do not  respond to PECS teaching in the classroom.  

Objectives :  (1) To investigate, in a pragmatic, group RCT the impact of providing PECS training and on-going monitoring to teachers on children’s communication skills.  (2)  To determine the characteristics of children who respond best to such training and the aspects of communication that are most likely to improve. (3) To consider the limitations and challenges of conducting treatment trials in the classroom- i.e. when therapy is administered directly by teachers rather than specifically trained therapists  

Methods :  84 children from 17 specialist elementary schools participated in the study. Mean age was 6.8 years (range 4-10 years); median non-verbal IQ was 29.9 (interquartile range 21.2-40.5). Over 80% had no spoken language or used single words only. All met criteria for autism or ASD on the ADOS, Class groups were randomly assigned to Treatment groups (immediate or delayed) or No Treatment. In the Treatment groups classroom staff attended 2-day workshops held by PECS consultants and were then regularly monitored and provided with feedback by expert PECS consultants over the following 5 months (6 sessions per month)  Multi-level Poisson regression was used to take account of initial group differences, treatment conditions, change across time points, and variations in length of observations and time between observations.

Results: Naturalistic observations of children in the classroom indicated that  PECS training for teachers resulted in significantly improved spontaneous communication using pictures, speech or both, in the children (p<.001). Spontaneous requesting for objects increased significantly (p<.001) but spontaneous requesting for social purposes did not. Only the effect on spontaneous speech persisted by 9 month follow-up. Improvement in the spontaneous use  of pictures to communicate was not associated with any specific child characteristics (e.g. DQ, initial language, ADOS score) , but  the children who showed improvements in speech tended to show less severe autism symptoms and to have at least some expressive language when intervention began.

Conclusions : Multi-level modelling allowed for exploration of specific intervention effects and outcome predictors. However, although it was possible to collect detailed information on child factors related to outcome, assessment of other variables, such as teacher competence, treatment fidelity, differences between classrooms or other factors related to school or home background was not possible within the design of the programme. Why treatment effects diminished over time is also unclear. The practical difficulties of conducting RCTs within real-life classroom settings will be further discussed.