Abstract
King, A. M., Hengst, J. A., & DeThorne, L. S. (2013). Severe speech sound disorders: An integrated multimodal intervention. Language, Speech, and Hearing Services in Schools, 44, 195–210.
Children with speech sound disorders (SSD) represent a large percentage of the caseloads of speech-language pathologists (SLPs). SSD include articulation disorders, phonological disorders, childhood apraxia of speech (CAS), and motor speech disorders. Traditional approaches for treating severe SSD rely on clinician-directed impairment-focused interventions that are designed to target a child’s accurate production of speech sounds. The goal of impairment-focused approaches is to remediate underlying impairments and improve individual abilities that are assumed to be necessary prerequisites for functional communication. Impairment-focused treatments for severe SSD, however, may not be feasible. They may yield too slow progress or change for some children, particularly those who are not easily stimulable or have difficulty imitating. In addition, impairment-focused approaches do not address the ultimate goal of intervention—improvement of children’s ability to communicate in natural settings. The International Classification of Functioning (World Health Organization, 2001), which serves as the foundation for clinical and research practice for SLPs, recommends two types of intervention goals: (a) impairment-based goals that address the skills needed to improve the child’s speech and language capacity so as to reduce restrictions in activity, and (b) social-based goals that address what the child needs to perform—to communicative effectively in the family, school, and community. To achieve social goals in cases of severe and persistent SSD, clinicians may introduce augmentative and alternative communication (AAC) systems to compensate for the child’s severely impaired natural speech. In contrast to impairment-focused approaches that are designed primarily to target speech sound production, AAC interventions are designed to bypass the activity limitations and participation restrictions of severe SSD, supporting children’s participation in social activities.
AAC approaches to intervention support communication by using multiple modalities, which contrasts with traditional treatments that target modality-specific skills and behaviors, such as removing the child’s AAC device when targeting speech production goals, insisting on AAC responses (instead of gestures or talk) when working on AAC mastery, and targeting spoken AAC and speech production goals through different activities. SLPs and parents often conceptualize AAC approaches as shifting treatment from natural speech production to alternative communication modalities. This move often is not made until after years of failed traditional speech treatment. In addition, parents, educators, and some SLPs share the concern that implementing AAC will interfere with children’s natural speech production abilities. Because of the impairment-focused approaches to SSD, concerns with the effect of AAC on natural speech development, and a lack of empirical evidence to show that AAC can effectively target natural speech, it is currently not common practice to incorporate AAC as an integral resource and treatment component to address children’s speech production goals.
Recent systematic reviews on natural speech development and AAC (e.g., Bauman Leech & Cress, 2011; Millar, Light, & Schlosser, 2006; Schlosser & Wendt, 2008) have suggested that AAC may have positive effects on the natural speech and language development of some children with severe communication impairments. Several possible learning mechanisms may be responsible for increases in natural speech through the use of AAC:
AAC permits persons to participate in social interactions in which learning takes place;
Speech-generating AAC may provide the scaffolding children to succeed in treatment;
A reduction in the pressure to speak (as a result of the use of AAC) might facilitate natural speech;
Speech-generating AAC systems may provide reinforcement for desired verbal behavior.
These hypotheses support use of an integrated multimodal intervention (IMI) that uses a variety of speech production and augmentative techniques to practice natural speech production within meaningful social communication contexts. The theoretical frameworks and empirical evidence suggest that the IMI will promote natural speech production in children with SSD who use speech-generating AAC by stimulating the communication system broadly. The purpose of this study was to examine the effects of an IMI that incorporated speech-generating AAC and traditional speech interventions, with the primary goal of increasing the quality of natural speech production of target speech sounds, and a secondary goal of increasing the quantity of natural and AAC speech in these children with severe SSD.
The researchers asked the following questions:
Primary research question:
Will children with severe SSD increase the quality or accuracy of their naturally produced target speech sounds following the IMI?
Secondary research questions:
Will children with severe SSD increase the quantity or amount of their naturally produced speech and AAC speech following the IMI?
Are increases in the quality of natural speech targets generalized to words containing the speech sound targets that were not a component of the IMI?
Are increases in the quality of natural speech maintained after the IMI is discontinued?
The Study
Participants
Three boys met the prerequisite criteria for the study, including (a) primary impairment of a severe SSD, (b) age between 4 and 9 years, (c) no report of hearing impairment, (d) primary form of communication was spoken English, and (e) currently using or had experience with a dynamic-display speech-generating AAC system. On the basis of the testing conducted at the preintervention assessment session, the three children also met the criterion of a severe/profound rating on either the Hodson Assessment of Phonological Patterns–3rd Edition (HAPP-3; Hodson, 2004)—or the Speech subtest of the Functional Communication Profile–Revised (FCP-R; Kleiman, 2003).
Method
This study used a multiple probe across participants design to evaluate the effectiveness of the IMI to increase the quantity and quality of word-level productions of target speech sounds for three participants with severe SSD. For each child, a set of 50 words was selected to be monitored or treated: 40 contained the participant’s target speech sound (10 baseline words, 20 treatment words, 10 generalization probe words) and 10 words contained the participant’s control speech sound (an error sound that was not treated). All children attended two to four baseline sessions and 9 to 12 IMI sessions. All sessions included three activities: (a) shared storybook reading, (b) natural speech target drill, and (c) structured play. Each boy was kept in treatment for a maximum of 12 IMI sessions, or until he had 90% accuracy on the target speech sound in treatment words during the three activities in the IMI session. In the baseline condition, the child’s AAC device was present and the research drew the child’s attention to it, but during the session made no effort to elicit AAC use, and used no augmented input. In the IMI treatment sessions, the researcher actively encouraged the child to use his AAC device to communicate.
The IMI consisted of five primary components:
Augmented input: the practice of the clinician accessing symbols on the AAC system the child is using, thus providing a spoken model and an AAC-generated model
Target redundancy: clinician provision of multiple meaningful models to naturally elicit participant productions
Naturalistic milieu language treatment techniques: including self-talk, parallel-talk, direct and gentle questioning, imitation, cloze techniques, expansions, recasts, and time delays
Correction procedures, use of placement cues, requests for imitation, and sound approximations intended to refine incorrect production of speech sounds
Results
Primary research question: Quality of natural speech
All three participants produced their treatment words more accurately during the IMI condition as compared with their baseline words during the baseline sessions. One boy met the accuracy criterion (90% over two consecutive sessions) before the 12th IMI session. Average percentage accuracy was also calculated for each participant in baseline and IMI conditions. One boy increased from 0% accuracy average in baseline to 55.3% accuracy average in the IMI, another boy increased from 0% to 73.5%, and the third boy increased from 13.2% to 53.4%. One boy also exhibited a significant increase in accurate production of his control probe words. Two boys exhibited no increase in accurate productions of the control probe words.
Secondary Research Question 1: Quantity of natural and AAC speech
Increases were observed in the total amount of natural speech produced in all the participants when compared with productions in the baseline sessions. There were smaller changes observed in the amount of AAC speech from the baseline condition to the IMI condition.
Secondary Research Question 2: Generalization of natural speech improvements
All the participants showed some amount of generalization to nontreated words that contained the target speech sound. For all participants, there was no generalization that occurred during the baseline condition. The data show that the generalization that did occur was higher when production was assessed through the picture-naming probe as opposed to spontaneous productions of nonstudy words during the sessions.
Secondary Research Question 3: Maintenance of natural speech improvements
Maintenance data were collected during the postintervention session that occurred 1 month following the final IMI session. All three participants maintained treatment word production accuracy levels that were similar to, or greater than, the accuracy levels that were obtained near the conclusion of the IMI. Moreover, results from readministration of the HAPP-3 indicated that all three participants decreased their overall phonological deviancy score and decreased the percentage of occurrence of the phonological processes that were targeted in the IMI.
All three parents interviewed provided positive feedback regarding the nature of the IMI. One parent commented that since the IMI, his son was verbalizing a lot more and was better at finishing his sounds, even in words that had not been practiced. He reported that because his son was using final consonants more, the family better understood his speech. The mother of another child reported that her son’s speech and communication were dramatically better and that 90% of the time, he communicated with his natural voice. She stated that it was the only thing she had seen that had that much success. One parent reported no change in the quantity of her son’s natural speech since the beginning of the IMI.
Comments
SLPs and parents are often apprehensive about implementing AAC with children with SSD due to concerns that AAC might inhibit or negatively affect their natural speech production. AAC interventions have historically been implemented to support activities and participation only after impairment-focused interventions failed. This study demonstrated that the integration of speech-generating AAC and traditional speech intervention techniques was not detrimental to natural speech, and in fact supported natural speech production on several levels. Natural speech was elicited during a meaningful variety of communication exchanges, not merely during a requesting or naming activity. The improvements in natural speech production accuracy are significant in two ways. First, for two participants, the improvements in speech accuracy were observed in the target speech treatment words, and there were no improvements observed in the production accuracy of the control words. The lack of improvement observed in the control words for two of the participants provides support for the internal validity of the study results. Data from postintervention HAPP-3 also support this claim. For two of the boys, although some decreases in phonological processes were evident in patterns that were not targeted during the intervention, the most significant decreases were evident in the targeted speech sounds. When considering data from these two levels of experimental control evident, the results are convincing that this intervention program led to significant increases in their ability to produce targeted speech sounds.
The boys in this study had received a diagnosis of CAS; they all imitated natural speech, and AAC was not their preferred mode of communication. The ability to imitate has been shown to be a significant predictor of natural speech development. Because all the participants were imitators, it is unclear whether this type of treatment would be effective with populations of children who did not have imitation skills. Although there is growing evidence for the value of using integrated and multimodal interventions to target children’s speech production, clinical practice still seems to prefer targeting isolated skills one at a time. This study offers a systemic approach that clinicians can adopt to communicatively engage clients in a clinical setting and simultaneously provide diverse and repeated opportunities for treating speech sound production. Success of the IMI in this study required the researcher (a) to carefully select activities and meaningful target words for each child; (b) to function flexibly as a communication partner throughout the session; (c) to take advantage of the repeated use of treatment words across activities to model and draw attention to target speech sounds; and (d) to modify the child’s productions of those sounds in meaningful words. In this way, the IMI components should be seen as a clinical tool kit that clinicians can adapt to use within meaningful activities and with salient words designed for specific clients.