Communication Partners Supporting Children With Complex Communication Needs Who Use AAC

Abstract

Communication partners who efficiently use augmentative and alternative communication (AAC) are essential interaction partners for children learning to communicate using AAC. This systematic review examines studies targeting interventions designed to help communication partners support children with complex communication needs who use AAC. Overall, the 13 studies of moderate methodological quality rated using published rating scales reported positive findings for partners’ skills and children’s communication. Author-reported effect sizes and calculation of improvement rate difference indicated that multiple studies demonstrated large effects that were maintained 1 to 2 months post intervention. Delivery considerations for partner training including device training, enhancing partners’ understanding of communication targets, and the timing of partner training are discussed. The importance of supporting successful interactions across learning contexts and domains of communication as well as over time is highlighted. Furthermore, implications for dyads with unique characteristics including limited experience with AAC and children who are developmentally young are examined.

Keywords

augmentative and alternative communication (AAC)communication intervention strategies development exceptionalities

The purpose of augmentative and alternative communication (AAC) is to “enable individuals to efficiently and effectively engage in a variety of interactions and participate in activities of their choice” (Beukelman & Mirenda, 2005, p. 8). Interventions for people using AAC often solely focus on an individual’s ability to navigate the AAC system, only one component of the bidirectional, co-constructed process of communication. To ensure that children with complex communication needs who use AAC have access to engage in the social world, it is essential to secure opportunities for children to engage with competent communication partners (Light & Drager, 2007). A communication partner could include a child’s interventionist, caregiver, or school personnel. Considering the viewpoint that language learning occurs within the context of social interactions between novice communicators and a partner (Tomasello, 2001), just like novice communicators who speak, children using AAC can also benefit from an environment that is infused with language models and symbols that are relevant to the child’s focus of attention by a partner (Brady, Herynk, & Fleming, 2010). Yet, the AAC system adds an extra dimension that complicates these social interactions creating a quadratic interaction between two people, a shared focus of attention, and the AAC system (Benigno & McCarthy, 2012). It is necessary to support communication partners’ ability to provide frequent and high-quality social interactions critical for children’s language development in this complex interaction. This systematic review will examine intervention studies designed to help partners learn to support children with complex communication needs who use AAC, focusing on the adult partners’ fidelity of intervention strategy implementation.

Interactions Including AAC

Both aided AAC systems (e.g., picture exchange communication system [PECS], speech-generating devices [SGDs]) and unaided AAC systems (e.g., gestures, manual sign, facial expressions) vary in their ease of use for a novice partner (Light & Drager, 2007). It may be challenging for partners who are not efficient AAC users to integrate AAC and provide the same frequency and quality of learning opportunities as partners using only spoken language (Brady et al., 2010). Broadly, communication partners require two types of skills. First, the partner must be able to understand the way the child is communicating, including spoken communication (SC), nonverbal communication as well as any mode of AAC used by the child. Second, the partner must also appropriately model augmented communication and respond to the child via this same array of types or modes of communication. Research to date suggests that the behavior of communication partners in these dyads may differ from spoken language dyads in important ways. For example, communication partners of children using AAC tend to dominate and interrupt interactions with question asking and otherwise redirection of the child’s attention leaving few communicative turns for the child (Kent-Walsh & McNaughton, 2005). This style of interaction restricts development to a limited range of communication functions and places the child who uses AAC in the passive role of a responder rather than an initiator of communication (Kent-Walsh & McNaughton, 2005). As such, partners may require strategies to interact with children using AAC in a style that facilitates both responding and initiating.

Training Communication Partners: Verbal and AAC Modalities

Research examining spoken language interventions demonstrates that communication partners can learn to facilitate children’s language development. Roberts and Kaiser’s (2011) meta-analysis showed that parents can successfully deliver interventions leading to gains in communication skills for children with speech and language delays. Moreover, clinicians in research settings have been shown to use strategies such as keeping the AAC system within reach, prompting communication, and using multiple stimuli to encourage AAC usage (Snell, Chen, & Hoover, 2006). Overall, many of these interventions follow principles of behavior focusing on antecedent, behavior, and consequence frameworks for teaching (Snell et al., 2006) and focus on participants with language delays as well as developmental disabilities.

Specific Aims of the Current Review

For the purpose of the review, the term communication partner will be used to refer to any individual engaged in communicative interactions with an individual communicating via an AAC system. In intervention studies, communication partners are those receiving support to enhance interactions with the child using AAC. Valuable narrative reviews have explored communication partner supports by focusing on theoretical instructional frameworks (Granlund, Bjorck-Akesson, Wilder, & Ylven, 2008) or by highlighting trends (Kent-Walsh & McNaughton, 2005); however, these reviews have not provided an extensive review of partners’ strategy implementation to facilitate children’s AAC use. Therefore, this systematic review will explore the efficacy of teaching practices designed to help communication partners learn to engage with children with complex communication needs who communicate using AAC systems. Due to the focus on partner training, only studies where partners’ strategy implementation is included as a dependent variable with numerical data have been included. Specifically, the review will:

identify experimental studies that have examined practices to help communication partners interact with children using AAC systems;

evaluate the methodological quality of the evidence using published rating scales; and

highlight considerations regarding the development and delivery of communication partner training programs.

Method

Search Strategy

A comprehensive search of 10 electronic databases spanning biomedical, educational, and medical content areas was completed in October 2012 and updated October 2013. Online First releases from relevant journals (e.g., Augmentative and Alternative Communication,; Autism; Journal of Speech, Language, and Hearing Research; Journal of Autism and Developmental Disorders) were scanned to pick up articles that may not yet be indexed by electronic databases. A detailed description of the search string, and the full set of searches can be obtained from the first author.

Study Selection

Gray literature including theses and dissertations was included, and no restrictions were placed on the date of publication. Only studies published in English were included. The inclusion criteria were designed to select experimental quantitative studies that provided empirical data for both partners and for children with complex communication needs who are using AAC. Studies included in this review met the following inclusion criteria: (a) Utilized a quantitative experimental design including group or single-subject experimental research designs (SSRDs). Studies providing quantitative data utilizing preexperimental designs (e.g., pre–post with no control group, case study designs) were not included due to the methodologically low level of evidence provided (Logan, Hickman, Harris, & Heriza, 2008); (b) Included children using an AAC system and who have complex communication needs. No restrictions were placed on the children’s diagnosis or the type of AAC system; (c) Included an intervention delivered to adult communication partners of children using AAC. Studies examining AAC interventions with no explicit teaching of partners were excluded. Peers were excluded in the definition of communication partners due to the differing nature of programming required for child peers versus adult partners; (d) Included communication partner fidelity or other partner behavior as a primary outcome (dependent variable) where numerical data for both partners’ and children’s outcomes were provided. Studies including outcomes only for the children who use AAC were excluded (e.g., Howlin, Gordon, Pasco, Wade, & Charman, 2007); (e) Included a primary outcome (dependent variable) in communication or language for children using AAC; (f) Studies using SSRDs must include graphical data to allow for calculation of improvement rate difference (IRD: Parker, Vannest, & Brown, 2009) for at least one primary outcome. Notably, three studies examining a reading-based intervention for children with Rett syndrome were excluded due to a lack of graphical data appropriate for IRD (e.g., Koppenhaver, Erickson, & Skotko, 2010); (g) Articles detailing a subset of a participant sample already included in another article were excluded (e.g., Romski, Sevcik, Adamson, Cheslock, & Smith, 2006 included 30 of 68 participants from Romski et al., 2010).

Assessment of Methodological Quality

Published methodological rating scales for randomized control trials (RCTs) and SSRDs were used to assess the quality of the studies. Based on recommendations made by the Cochrane Collaboration (Higgins & Green, 2011), the Jadad et al. (1996) scale was used to evaluate the RCT. The seven-item Jadad scale focuses on the processes and documentation of randomization and blind assessment and provides a quality rating score of “low” (0–2 points) or “high” (3–5 points). A limitation of the scale is that it demands double blind assessment for a high rating, which is not always appropriate for intervention trials where caregivers or educational personnel are receiving training. Given this limitation, an additional seven-item scale covering items related to participant inclusion criteria, adherence to intervention assignment, validity of measures, blind assessment, appropriate statistical analyses, attrition, and control for confounds was included (American Academy for Cerebral Palsy and Developmental Medicine [AACPDM], 2008). This scale provides an overall quality rating of “weak” (0–3 points), “moderate” (4–5), or “strong” (6–7).

Studies using SSRDs were assessed using a 14-item scale by Logan et al. (2008) and a supplementary 7-item scale (Smith et al., 2007), which captured important methodological items that are not found in the AACPDM scale (e.g., maintenance/generalization and intervention fidelity). The Logan et al. (2008) scale provides an overall quality score of “strong” (11–14 points), “moderate,” (7–10), or “weak” (0–6). No summary ratings are available for the Smith et al. (2007) scale; therefore, the total number of points achieved out of a total of 7 possible items was reported for each study (see Table 2). These scales have been used in prior published work with high reliability (e.g., Patterson, Smith, & Mirenda, 2012).

All studies were assessed by two independent raters. Intra-class correlations (ICCs) were calculated for each scale, indicating high reliability for the single-subject scales (Logan et al., 2008: α = .821; Smith et al., 2007: α = .954). Due to the small number of items, ICCs could not be calculated for the group study (Romski et al., 2010), so inter-rater agreement was calculated by dividing agreements by disagreements for each item. Agreement was 100% for the Jadad et al. (1996) scale and 83.3% for the AACPDM (2008) scale. Following Cochrane Collaboration standards, discrepancies were resolved through discussion with a third independent rater (Higgins & Green, 2011).

Data Extraction and Analysis

Data encompassing participant characteristics, research design, intervention, context, outcomes, statistical and/or visual analysis, and study conclusions were extracted by one reviewer using a pretested data extraction form adapted from Ospina et al. (2008). A second reviewer verified the accuracy and completeness of the extracted data.

Author-reported effect sizes for group designs are described, while IRD (Parker et al., 2009) was calculated for studies using SSRDs. IRD has a strong history in medicine and allows for the calculation of the magnitude of the intervention effect in SSRDs (Parker et al., 2009). IRD has the advantages of the ability to obtain confidence intervals, no data assumption required for use, hand calculation, and ease of interpretation (Parker et al., 2009). As such, IRD was selected for estimation of overlapping data.

Results

A systematic search of the literature targeting training programs for partners working with children with complex communication needs who use an AAC system led to 3,994 citations. Citation titles and abstracts were examined for any indication that the study systematically addressed partner training. Five hundred sixty-seven of those articles were identified as relevant and thus retrieved and examined against the inclusion criteria. Eleven papers were included. An updated search produced an additional 1,009 citations where 27 manuscripts were retrieved and examined leading to the inclusion of two papers for a total of 13 that met all seven inclusion criteria.

Participant Characteristics

Communication partners

Parents (87 mothers, 8 fathers) were included as partners in 10 studies (see Table 1). Parents’ ethnicity included African American (n = 19), Asian (n = 5), Caucasian (n = 39), and Hispanic (n = 5) as reported for 68 parents (Binger, Kent-Walsh, Berens, Del Campo, & Rivera, 2008; Nunes & Hanline, 2007; Romski et al., 2010; Rosa-Lugo & Kent-Walsh, 2008). Parents’ level of education as reported for 72 parents included high school degree (n = 9), completion of some college (n = 11) or a college degree (n = 30), an associate’s degree (n = 1), enrollment or completion of a post-graduate degree (n = 21; Binger et al., 2008; Cafiero, 1995; Kent-Walsh, Binger, & Hasham, 2010; Nunes & Hanline, 2007; Romski et al., 2010; Rosa-Lugo & Kent-Walsh, 2008).

Table 1.

Communication Partner Information and Outcomes.

Study	Communication partner	Intervention	Primary outcome	Reported effect size (group designs)IRD and 85% confidence interval (SSRDs)
Binger, Kent-Walsh, Berens, Del Campo, and Rivera (2008)	Parent: Female (2), male (1); some college/degree (3); Hispanic (3)	Book reading activity	Percentage of steps correctly implemented	Intervention—1 large; Generalization—1 large; Maintenance—1 large
Binger, Kent-Walsh, Ewing, and Taylor (2010)	Educational assistants (3): females; some college/degree (3); African American (AA: 1), Hispanic (2)	Book reading	Percentage of steps correctly implemented	1 large
Bingham, Spooner, and Browder (2007)	Educational assistants (3): female; high school (3); experience: 2.5–65 months	Classroom AAC use	Number of prompts and responses to child	Prompting child: 1 large Responses to child: 1 large
Cafiero (1995)	Parent: HS (1), college degree (1), education NR (1); lower middle class (2), upper class (2), NR (1)	Natural aided language stimulation	Intervals with: PCS; parent–child talk	PCS (n = 4): 0.73 [0.62, 0.84] moderate; Parent communication (n = 4): 0.14 [0.12, 0.16] small; Stress: No graphical data for IRD
Chang (2009)	Parent: No other information provided	Mass trials	Average implementation score	After 1 week training: 0.23 [0.20, 0.26] small After 6 week training: 0.28 [0.24, 0.32] small
Douglas, Light, and McNaughton (2012)	Paraprofessionals (3): female (2), male (1), college degree (2), nurse (1)	IPLAN and MORE	Communication opportunities provided	Intervention—1 large Maintenance—0.93 large
Ganz et al. (2013)	Preschool therapists (3): Female (2), male (1), master’s degree (3); experience: 3–4 years (2), 4 months (1)	PECS	Frequency of PECS opportunities provided	Intervention—1 large Generalization—0.17 small
Kent-Walsh, Binger, and Hasham (2010)	Parents: Female (6), high school (1), some college/degree (4), associate’s degree (1); AA (3), Caucasian (3)	Book reading	Level and rate of parent acquisition of skills	No graphical data provided
Iacono, Chan, and Waring (1998)	Parent; females; middle-upper class (5)	Milieu	Parent utterance type	No graphical data provided
Nunes and Hanline (2007)	Parent: Female; high school (1); low socioeconomic status (1)	Four naturalistic teaching strategies	Environmental arrangement, Mands, Mands with AAC, and Models across PR and CR	Environmental arrangement: PR: 0.88 large; Gen: −0.66 none; CR: 0.47 small; Gen: 0 none Mands: PR: 0.25 small; Gen: 0.33 small CR: −0.22 none; Gen: 1 large Mands with AAC: PR: 0.58 moderate; Gen: −0.33 none; CR: −0.11 none; Gen: −0.33 none Models: PR: 0.25 small; Gen: 0 none; CR: −0.11 none; Gen: 0 none
Romski et al. (2010) RCT	Parent: High school (6), some college or degree (33), post-grad (21), no report (2); AA (18), Asian (5), Caucasian (39)	Augmented and spoken strategies	Parents: Mean length utterance, mean length of turn, total turns	Group by session (three conditions) effect size: MLU: $η_{p}^{2}$ = .02 NS; mean length turn: $η_{p}^{2}$ = .17 NS; parent total turns: $η_{p}^{2}$ = .27 small
Rosa-Lugo and Kent-Walsh (2008)	Parent: Female (2), high school (2); Hispanic (2)	Book reading	Accuracy of implementation	Intervention—1 large; Generalization—1 large; Maintenance—1 large
Stiebel (1999)	Parent: No other information provided	Problem solve	Chances to communicate	0.96 (0.82–1.10) large

Note. IRD = improvement rate difference; SSRDs = single-subject experimental designs; AAC = augmentative and alternative communication; HS = high school; NR = not reported; PCS = picture communication symbols; IPLAN = Identify activities for communication, Provide means for communication, Locate and provide vocabulary, Arrange environment, use iNteraction strategies; MORE = Model AAC, Offer opportunities for communication, Respond to communication, Extend communication; PECS = picture exchange communication system; PR = play routines; CR = caregiver routines; RCT = randomized control trial; MLU = mean length of utterance; NS = not significant.

Educational assistants were included in three studies (see Table 1). They were primarily female (8 of 9); had completed high school (n = 3), some college (n =2) or a college degree (n = 4); and were of a range of ethnicities (African American = 1; Hispanic = 2; unknown: 6). All assistants had 3 to 5 years of classroom experience except two who had less than 4 months.

Children using AAC

Studies enrolled a total of 105 participants (see Table 2 for characteristics). The majority of participants were male (n = 72). Ethnicity as reported for 76 participants across six studies was Caucasian (n = 40), African American (n = 23), Asian (n = 7), and Hispanic (n = 6). Studies included 87 toddlers or preschoolers, 17 school-age children, and 1 adolescent. The children received a range of diagnoses, including autism or autism spectrum disorder (n = 16), pervasive developmental disorder not otherwise specified (n = 2), bilateral schizencephaly (n = 1), cerebral palsy (n = 3), cleft palate (n = 1), cystic hygroma (n = 1), DiGeorge syndrome (n = 1), Down syndrome (n = 7), dysarthria (n = 1), phonological processing disorder (n = 1), and unspecified developmental delays (n = 9). One study (Romski et al., 2010) including 62 of the 105 total participants provided only a general description of diagnoses received by participants, including genetic syndromes, seizure disorders, and cerebral palsy.

Table 2.

Child Participant Information, Study Outcomes, and Methodological Quality.

Study and design	N, Age, Gender, Ethnicity	Diagnosis	Intervention and primary outcome	Type of AAC, prior AAC experience and language level, Range (median)	Primary outcomes: reported effect size and IRD	Methodological quality
SSRDs
Binger, Kent-Walsh, Berens, Del Campo, and Rivera (2008) Multiple probe	N = 3 Preschool (3) Male (1), female (2) Hispanic (3)	Phonological processing disorder (1), DiGeorge syndrome (1), cleft palate (1)	Book reading activity (NR) and frequency of spontaneous and imitated messages	Communication board (1) SGD (2) Yes (1) TACL-3 SS 106–109 (106)	Intervention—1 large Generalization—1 large Maintenance—1 large	Logan et al. (9); Smith et al. (5) Moderate
Binger, Kent-Walsh, Ewing, and Taylor (2010) Multiple probe	N = 3 Preschool (1), school age (2) Male (2), female (1) African American (2), Hispanic (1)	Developmental delay (2), dysarthria (1)	Book reading (3–5 × 10-min sessions) and frequency of multi-symbol messages	SGD (3) Not reported TACL-3 SS/AE 61–98 (85)/3;6–5;7 (3;6)	Intervention—1 large Generalization—1 large Maintenance—1 large	Logan et al. (11); Smith et al. (5) Strong
Bingham, Spooner, and Browder (2007) Multiple probe	N = 3 School age (2), adolescent (1) Male (1), female (2)	Developmental delay (3)	Classroom AAC use (1–2 hr intro, 6 hr role playing, 2 hr on self-evaluation) and frequency of responding with AAC; frequency of problem behavior	Picture symbols (3) Big Macs (1) Must use an AAC system to be included Baseline AAC frequency 0–1 (0)	Responding with AAC: Intervention—0.78 moderate Follow-up—1 large Occurrences of problem behavior: Intervention—0.59 moderate Follow-up—1 large	Logan et al. (3); Smith et al. (3) Weak
Cafiero (1995) Multiple probe	N = 5 Preschool (5) Male (4), female (1)	Autism (4), PDD (1)	Natural aided language stimulation (2 × group training, 7–10 hr in home) and child responses, initiations with picture communication symbols	Picture symbols (3) Used AAC for at least 6 months No assessments reported Parents reported: babbling through 1 to 3 word utterances	Child communication (n = 4): 0.31 small Parent–child communication (n = 4): 0.44 small	Logan et al. (2); Smith et al. (3) Weak
Chang (2009) Multiple baseline	N = 6 Preschool (1), school age (5) Male (6)	Autism (6)	Mass trials and parent training targeting responses across three routines (e.g., play, snack) for each participant	Picture cards (6) No past AAC exposure VABS communication AE 0;1–0;8 (0;5)	Across 3 routines: Phase 1: 0.66–0.78 moderate, Phase 2: 0.58–0.98 moderate-large Parent training: 0.58–1 moderate-large	Logan et al. (6); Smith et al. (3) Weak
Douglas, Light, and McNaughton (2012) Multiple baseline probe	N = 3 Preschool (3) Male (1), female (2)	Down syndrome (1), developmental delay and hard of hearing (1), bilateral, schizencephaly (1)	IPLAN and MORE intervention strategies in 15-min play sessions	Voice output (1); sign and gesture (2) Some anecdotal exposure to sign (children using 1–3 single signs)	Child communication turns: Intervention—1 large, Maintenance—0.80 large	Logan et al. (8); Smith et al. (3) Moderate
Ganz et al. (2013) Multiple probe	N = 3 Preschool (3) Male (2), female (1)	Autism (1), ASD (1), PDD-NOS (1)	PECS	Children have mastered at least one phase of the PECS protocol	Independent PECS exchanges: Intervention—0.76 moderate Generalization—0.25 small	Logan et al. (12); Smith et al. (4) Strong
Kent-Walsh, Binger, and Hasham (2010) Multiple probe	N = 6 Preschool (1), school age (5) Male (4), female (2) African American (3), Caucasian (3)	Cerebral palsy (3), Down syndrome (3)	Book reading (2–2.5 hr) and number of communication turns (comments and questions) via any modality; different concepts	Communication board (1) Manual sign (1) SGD (4) At least 10 AAC symbols Peabody Picture Vocabulary Test–III SS/AE 40–110 (62)/2;1-6;0 (2;5) TACL-3 SS/AE 49–104 (62)/3;0–5;2 (3;7)	Communication turns: Intervention—1 large, Generalization—1 large, Maintenance—1 large Different concepts: Intervention—1 large, Generalization—1 large, Maintenance—1 large	Logan et al. (11); Smith et al. (5) Strong
Iacono, Chan, and Waring (1998) Multiple baseline	N = 5 Preschool (5) Male (3), female (2)	Developmental delay (2), Down syndrome (3)	Problem-solving using milieu teaching (28 sessions at 2.5 hr each) and number of linguistic units; frequency of vocalizations and gestures; modality of linguistic unit	Manual sign (5) History not reported Reynell AE 1;8–2;0 (1;9) MCDI receptive total 206–279 (215) MCDI expressive total 14–72 (20)	Linguistic unit production (n = 5): 0.35 small Frequency of vocalizations (n = 2): 0.96 large Frequency of gestures (n = 2): 0.42 small Modalities (no graphical data)	Logan et al. (5); Smith et al. (1) Weak
Nunes and Hanline (2007) Multiple baseline	N = 1 Preschool (1) Male (1)	Autism (1)	Four naturalistic teaching strategies (2 hr per play and caregiver routine) and AAC use; gesture and manual sign; vocalizations and verbalizations (all variables in both PR and CR)	Communication board (1) History not reported LAP AE 21 months	Communication turns PR: 1 large, generalization (Gen): 1 large; CR: 0.09 small; Gen: 1 large Imitative response PR: 0.11 small; Gen: 0 None CR: 0 none; Gen: 0 none Child response AAC use PR 0.88 large; Gen: 0 none CR: 0.4 small; Gen: 0 none Verbal response—PR: 0.5 moderate; Gen: 0 none; CR: −0.11 none; Gen: 0 none Gesture and sign response: PR: 0.63 moderate; Gen: 1 large; CR: 0.2 small; Gen: 1 large	Logan et al. (8); Smith et al. (3) Moderate
Rosa-Lugo and Kent-Walsh (2008) Multiple baseline	N = 2 School age (2) Male (1), female (1)	Cystichygroma (1), developmental delay (1)	Book reading (1 hr intro, 2–5 hr practice) and number of communicative turns	SGD (2) At least 10 AAC symbols Receptive One Word Picture Vocabulary Test—SS: 55 TACL-3 SS 64–66 (65)	Intervention—1 large Generalization—1 large Maintenance—1 large	Logan et al. (6); Smith et al. (4) Weak
Stiebel (1999) Multiple baseline	N = 3 Preschool (2), school age (1) Male (3) Hispanic (2)	Autism (3)	Problem-solving home routines (1–2 × 1–1.5 hr) and frequency of spontaneous card use	Picture cards (3) History not reported Vineland Adaptive Behavior Scale: Communication-SS: 34–45 (38)	Intervention—0.63 moderate Maintenance—0.81 large	Logan et al. (8); Smith et al. (3) Moderate
Randomized control trials
Romski et al. (2010)	N = 62 Preschool Males (43), females (19)	Unspecified genetic and developmental disorders (autism excluded)	Naturalistic strategies in three conditions: ACI, ACO, and SC; 18 × 30 min) and number of augmented words and spoken words; number of different spoken words and vocabulary size; MLU; mean turn length; type/token ratio; utterance intelligibility; utterance rate; total turns	SGDs (62) History not reported Scores reported for AC-I/ AC-O /SC groups MSEL receptive SS 28/29/27; expressive SS 22/22/21 SICD receptive AE 18/20/19; expressive AE 12/13/13 MCDI receptive total score 114/144/155; expressive total score 11/14/8	Group by session Augmented words Intervention: ACO > ACI $η_{G}^{2}$ = .11 NS Spoken words $η_{G}^{2}$ = .04 NS Vocabulary size η² = .47 ACO > ACI > SC—Significant post hoc test MLU $η_{p}^{2}$ = .15, mean turn length $η_{p}^{2}$ = .14 Type/token ratio $η_{p}^{2}$ = .26 Intelligibility $η_{p}^{2}$ = .59, utterance rate $η_{p}^{2}$ = .26 Total turns $η_{p}^{2}$ = .26	Jadad et al. (1); AACPDM (5) Low/moderate

Note. AAC = augmentative and alternative communication; IRD = improvement rate difference; SSRDs = single-subject experimental designs; SGD = speech-generating device; TACL = Test for Auditory Comprehension of Language; SS/AE = Standard Score/Age Equivalent; PDD = pervasive developmental disorder; VABS = Vineland Adaptive Behaviour Scale; IPLAN = Identify activities for communication, Provide means for communication, Locate and provide vocabulary, Arrange environment, use iNteraction strategies; MORE = Model AAC, Offer opportunities for communication, Respond to communication, Extend communication; ASD = autism spectrum disorder; PDD-NOS = pervasive developmental disorder not otherwise specified; PECS = picture exchange communication system; MCDI = MacArthur-Bates Communication Development Inventory; PR = play routines; CR = caregiver routines; LAP = Learning Accomplishment Profile; ACI = Augmentative Communication Input; ACO = Augmentative Communication Output; SC = spoken communication; MLU = mean length of utterance; MSEL = Mullen Scales of Early Learning; SCID = Sequenced Inventory of Communication Development; AACPDM = American Academy for Cerebral Palsy and Developmental Medicine.

Language skills

Table 2 outlines the language assessments (clinician delivered and standardized parent report measures) used in the nine studies reporting entry language. Across measures, the majority of participants had very limited language skills for their chronological age. Age equivalent receptive and/or expressive language scores from four studies (see Table 2) indicated language scores below 24 months (range = 1–24 months). In addition, in two studies (Iacono, Chan, & Waring, 1998; Romski et al., 2010) 73 of 75 participants entered with 20 or fewer expressive words (range: 8–20) based on raw MacArthur-Bates Communication Development Inventory (MCDI; Fenson et al., 1993) scores. Finally, four studies (Binger, Kent-Walsh, Ewing, & Taylor, 2010; Kent-Walsh et al., 2010; Rosa-Lugo & Kent-Walsh, 2008) indicated significant delays in receptive language, including eight participants at or below the 3rd percentile and three participants between the 16th and 61st percentiles (Test of Auditory Comprehension–3; Carrow-Woolfolk, 1999). Altogether, these nine studies included children with significant language delays.

AAC systems and exposure

Eleven studies examined aided AAC systems, whereas three included unaided AAC (e.g., manual sign and gesture) for at least one child (see Table 2). Aided AAC included communication boards or picture cards (eight studies) and/or SGDs (nine studies). Four studies failed to report children’s prior AAC exposure (Iacono et al., 1998; Nunes & Hanline, 2007; Romski et al., 2010; Stiebel, 1999). Seven children in two studies (Binger et al., 2008; Chang, 2009) had no AAC history. All other studies required AAC exposure based on duration (e.g., at least 6 months; Cafiero, 1995), number of symbols (e.g., at least 10; Rosa-Lugo & Kent-Walsh, 2008), or PECS mastery (Ganz et al., 2013).

Intervention: Protocol Types

Studies used a variety of strategies and contexts to teach partners to support children’s use of AAC (see Table 1). Intervention fidelity was measured in five studies, whereas individual strategies were examined in seven studies. Strategies included a range of support levels from minimal direct support such as modeling and environmental arrangement (e.g., Nunes & Hanline, 2007) to direct approaches such as verbal and physical prompting (e.g., Bingham, Spooner, & Browder, 2007). The majority of studies (8 of 13) took place in family homes. Communication skills were targeted in the context of academic or play settings (eight studies), home (two studies), and recreational routines (one study). Only three studies examined two or more contexts (see Table 1).

Book reading

In four studies (Binger et al., 2008; Binger et al., 2010; Kent-Walsh et al., 2010; Rosa-Lugo & Kent-Walsh, 2008), varied adaptations of a book reading intervention were examined as the context for increasing communication via AAC. The protocol varied by study but primarily included brief practice and feedback (e.g., 4–6 hr; Rosa-Lugo & Kent-Walsh, 2008) using either minimal supports (e.g., modeling, contingent responding, waiting, and open-ended questions) or a heavily prompted question-asking protocol focused on responding (e.g., Binger et al., 2010).

Other protocols

Novel protocols referred to as “naturalistic” (Cafiero, 1995; Nunes & Hanline, 2007) consisted of modeling communication behaviors as well as prompting and environmental supports. Both Douglas, Light, and McNaughton (2012) and Ganz et al. (2013) used play routines as the context to elicit communication opportunities. In addition, Chang (2009) used a mass trial learning approach (repeated sets of teaching trials), whereas Stiebel (1999) applied a home routine problem-solving intervention.

Comparison of conditions

Only one study compared multiple teaching conditions (Romski et al., 2010). Parent–child dyads were randomized to one of three conditions, which differed in two ways: (a) target output (AAC or SC) and (b) presence or absence of demand for output (AC output [ACO] or AC input [ACI]). Treatment consisted of 18 clinic sessions and 6 home sessions to teach parents prompting and environmental strategies.

Measures and Primary Target Outcomes

Communication partners

Fidelity of intervention implementation and strategy use were the most common partner outcomes (seven studies; see Table 1) as well as provision of opportunities for child communication (two studies). In addition, three studies examined partners’ mean length of utterance (MLU) as well as mean length and number of turns (Romski et al., 2010), the function and type of partners’ utterances (Iacono et al., 1998) as well as their style of interaction (e.g., prompting vs. contingent responding; Bingham et al., 2007). Cafiero (1995) explored parent stress using the Parent Stress Index (PSI; Abidin, 1995).

Children using AAC

Language outcomes obtained through live coding or coding of videotape included MLU, AAC use, gestures, vocalizations, and picture exchange. These outcomes were presented as one variable that collapses both responses and initiations. In addition, six studies examined transcribed interactions to explore expressive communication (e.g., spoken and augmented words, MLU, number and length of turns, communication rate). In addition, one study included frequency of problem behavior (Bingham et al., 2007).

Research Design

Studies included one RCT (Romski et al., 2010) using a three-group comparison design, and 12 SSRDs including multiple baseline and multiple probe designs.

Quality of Conduct

Quality of conduct scores for each study is described in Table 2. Itemized scores for SSRD quality are found in Table 3 and scores for the one RCT are described in the text to follow. The bulk of the included studies used SSRDs. These 12 studies varied in quality (3–12 points, median = 8) on the Logan et al. (2008) scale, including studies rated as “weak” (five studies), “moderate” (four studies), and “strong” (three studies). Corresponding scores on the Smith et al. (2007) scale ranged from 1 to 5 out of 7 (median = 3). Overall, interventions and outcomes were well described, research designs were accurately reported, and graphical data were accurately represented. Approximately half of the studies established stable baseline measures, demonstrated intervention effects across at least three participants, and reported generalization measures for all participants. However, the majority of studies failed to require blind assessment or measure treatment fidelity. Furthermore, the studies failed to adequately report participant characteristics or describe level, trend, or variability within the data. Finally, the RCT (Romski et al., 2010) was of low–moderate quality, scoring 2 points (low) on the Jadad et al. (1996) scale due to a lack of description of the randomization process and lack of double blind procedures. The study earned 5 points (moderate) on the AACPDM (2008) scale, due to well-described inclusion criteria, use of reliable measures, inclusion of power calculations, and controls for potential confounds.

Table 3.

Methodological Characteristics of Single-Subject Studies.

American Academy for Cerebral Palsy and Developmental Medicine (2008) quality components	Number of studies (%) N = 12
1. The participant(s) are well described to allow comparison with other studies or with the reader’s own patient population	5 (42%)
2. Independent variables were operationally defined to allow replication	8 (67%)
3. Intervention conditions operationally defined to allow replication	10 (83%)
4. Dependent variables were operationally defined	8 (67%)
5. Inter-rater/intra-rater reliability of dependent measures was assessed before and during each phase of the study	10 (83%)
6. Outcomes assessors were blind	0
7. Baseline data were stable	6 (50%)
8. Type of SSRD was clearly and correctly stated	12 (100%)
9. Adequate number of data points in each phase (min 5)	2 (17%)
10. Intervention was replicated across three or more subjects	6 (50%)
11. Authors conducted and reported visual analysis	4 (33%)
12. Graphs used for visual analysis follow standard conventions	8 (67%)
13. Authors report tests of statistical analysis	5 (42%)
14. All criteria met for statistical analysis	5 (42%)
Smith et al. (2007) quality components
1. Use of single-case experimental design	12 (100%)
2. Specific inclusion and exclusion criteria for enrollment in study along with dropouts and intervention failures	6 (50%)
3. Participant samples were well-defined (i.e., standardized diagnostic tests, standardized intelligence tests, adaptive behavior)	0
4. Intervention effects were replicated across three or more participants	6 (50%)
5. Assessment of generalization of intervention effects was conducted in at least one other setting or the study measured maintenance of treatment effects over time	10 (83%)
6. Measurement of outcomes was conducted by assessors who were blind to purpose of the study	0
7. Fidelity of intervention implementation was monitored through direct observation	8 (67%)

Note. SSRD = single-subject experimental design.

Outcomes

Author-reported effect sizes are shown in Tables 2 and 3 and interpreted within Cohen’s (1992) framework of small (.20–.49), medium (.50–.79), and large effects (.80–1.0). For studies using SSRDs, IRD analysis was conducted. IRD benchmarks (Parker et al., 2009) include small (0–.50), moderate (.51–.70), and large effects (.71–1.0).

Communication partner outcomes

From those studies that measured intervention fidelity and provided graphical data, it was observed that several partners began intervention with scores below 25% (e.g., Rosa-Lugo & Kent-Walsh, 2008), many near zero (e.g., Binger et al., 2008). Large effects of intervention were found on partner’s strategy use in three of four reading studies (see Table 1) and at follow-up, while small effects on fidelity of mass trial teaching were reported (Chang, 2009). Furthermore, when fidelity of a specific strategy was examined (e.g., prompting, responding, providing opportunities, modeling, and environmental supports), partners also entered intervention with near zero scores (e.g., Bingham et al., 2007; Ganz et al., 2013). Large effects were found for both partner prompting and provision of communication opportunities in two studies (Bingham et al., 2007; Douglas et al., 2012), while Nunes and Hanline (2007) reported mixed effects.

In addition, five studies examined partners’ discrete communication behaviors (e.g., partner’s MLU; Romski et al., 2010). Three studies found large increases in child communication opportunities provided by the partner (Douglas et al., 2012; Ganz et al., 2013; Stiebel, 1999), while Cafiero (1995) found moderate effects for partner communication frequency. Finally, Romski et al. (2010) found small effects for communication turn frequency and no effects for parent MLU or mean turn length.

Children’s outcomes

Overall, children made gains in AAC use, turns, and receptive vocabulary although some mixed effects were found. Effects were found across multiple studies for low-rate communicators, including children with zero functional symbols at entry who exited intervention with 2 to 15 symbols per 10- to 30-min session (e.g., Binger et al., 2008, Romski et al., 2010, Rosa-Lugo & Kent-Walsh, 2008). Examination of the intervention effects indicates large effects for children’s use of AAC in eight studies (see Table 1).

Spoken language

Spoken language and/or vocalizations were targeted in three studies. Large (Iacono et al., 1998) to moderate effects (Nunes & Hanline, 2007) on vocalizations were found. In the RCT (Romski et al., 2010), although some children showed increases, effects on spoken language were not significant. Romski et al. (2010) examined additional aspects of the interactions, finding a number of small and moderate effects for vocabulary and intelligibility where children using AAC experienced the greatest increases.

Follow-up and generalization

Large effects were also found for follow-up measures in six studies (see Table 2) for child communication outcomes, including messages during book reading and children’s communication turns. However, Nunes and Hanline (2007) found mixed generalization effects with large effects in play for gesture, sign, and turns, but no effects for AAC or spoken responding, and imitation.

Discussion

The importance of partner training has been highlighted in the extant literature, yet few empirical examinations of training programs for partners have been published. The 13 studies included in this review provide preliminary evidence of moderate methodological quality, demonstrating that both parents and educational assistants are adopting interaction strategies through brief and targeted instructional programs that positively influenced children’s AAC use.

Partners’ adoption and accurate implementation of communicative strategies and their impact on children’s outcomes may be influenced by a number of factors that need to be addressed for a full understanding of the efficacy of partner training programs. The studies reviewed here reflect some key limitations of the current literature and highlight critical methodological considerations for further research in this area. In particular, we will discuss the relevance of device training, timing of partner training, selection of developmentally appropriate communication targets for the child, and the maintenance and generalization of partners’ strategy implementation.

Device Training

What is missing from these training programs is explicit reporting of protocols used to help partners navigate the AAC systems including specific information about the baseline level of AAC proficiency of participants as well as methods to familiarize participants with the selected mode of AAC. All but two studies included children who had some experience with AAC, but no study reported on partners’ previous AAC history. Particularly for dyads that lack this experience, both parties may require training in how to use the AAC system prior to interaction training. Multiple studies included SGDs which require that partners be able to program new symbols and navigate the device features to effectively engage the child. Therefore, consideration of the type of AAC system and the breadth and depth of training needed is critical to make a fair comparison of the effectiveness of a training program across different AAC systems. Furthermore, partner training dosage and intervention effects varied across studies with similar intervention protocols but varied populations (e.g., Binger et al., 2008, Bingham et al., 2007). Without knowing the baseline-level AAC experience of the partners, it is difficult to draw conclusions about optimal intervention dosage from these studies. Further research should include in their designs an identification of the level of expertise of both the children and the partner relative to the type of AAC system to be used.

Timing the Introduction of Partner Training

The timing of the introduction of partner training also must be considered along with the level of experience with AAC systems. Two studies staggered the start of clinician–child intervention and partner training, starting with clinician only intervention before adding partner sessions (Chang, 2009; Romski et al., 2010). These studies included children with minimal or no AAC experience prior to intervention. Beginning with direct clinician–child intervention may have served to jump start the child’s use of the AAC. Pushing back the entry of partners into the interaction may allow both parties time to absorb basic strategies. Further examination of the timing of this merger with examination of partner fidelity and subsequent effects on children’s communication is required.

Teaching Targets

In addition, explicitly teaching partners the difference between children’s initiations of communication and responding to adult bids is crucial. Much like interventions targeting spoken language, spontaneous communication is an important AAC outcome (Mirenda, 2003). Children who rely on prompting to communicate (responding) would not meet this benchmark. Only four included studies targeted nonimitated and nonprompted spontaneous communicative initiations and collected these data separate from elicited (prompted) language. Together, these studies provide preliminary evidence for interventions targeting spontaneous communication in play and book-based interactions with parents (Binger et al., 2008; Kent-Walsh et al., 2010; Romski et al., 2010; Stiebel, 1999). The limited focus on spontaneous initiations and emphasis on responding in the literature has implications for how and what we teach partners to do to support children using AAC. Natural social interactions include both responses and initiations and as such, it is crucial to present a range of strategies to support both sets of skills.

Another consideration for target communication outcomes is the function of the bid. For example, social communication includes two broad categories, including requesting (e.g., asking for an object) and commenting (e.g., to share social interest in an activity). Although requesting is a frequent target in both communication interventions targeting spoken and augmented language, other social functions including commenting are very limited (Kasari et al., 2014). In addition to supporting children’s communicative initiations, fostering social initiations will enhance the child’s ability to engage in more flexible social interactions. It is notable that no study included in this review clearly or systematically documented the functions of the children’s spoken and augmented utterances. However, several studies did state within the text that comments were likely included in counts such as number of different words and number of utterances. For example, sample interactions provided by Romski et al. (2010) indicate that both prompted requests and prompted comments were included in the outcome variables. Therefore, it is critical that future studies not only explicitly differentiate and record spontaneous and prompted communication but also the function of those communicative bids.

Across Contexts and Over Time

Helping partners support children using AAC across multiple contexts and communication domains is also essential to securing high-quality opportunities for children’s language learning. The inclusion of multiple environments (e.g., home, community, school) and multiple contexts (e.g., play, homework, and home routines) within partner training protocols may increase partners’ ability and confidence to consistently use interaction strategies throughout the child’s day. Only 2 studies included more than one context with 11 studies focusing on one specific context such as book reading or play. Large effects were found for partners’ strategy use as well as children’s communication for book reading in four studies; but outside this context, the magnitude of effects varied widely. This lack of generalization across contexts was captured and led to a lower score on the Smith et al. (2007) scale. Future research targeting multiple contexts in both intervention and generalization programming is necessary to understand the degree to which children using AAC and their partners can flexibly engage in communicative interactions across the child’s day.

In addition, it is also important that partners are able to implement the strategies over time once consultation and support from the trainer has faded. Maintenance of intervention gains was measured in 8 of the 11 studies (see Table 3). Follow-up periods were brief (1–2 months), demonstrating varied continuation of treatment gains for both partners and children using AAC. Maintenance measures were taken in a context very similar to the intervention context with little variation in the materials used (e.g., different books; Kent-Walsh et al., 2010). Overall, mixed maintenance of treatment gains was demonstrated 1 to 2 months post exit.

Variability in Dyads: Children’s Developmental Level and Receptive Language

As communicative interactions are bidirectional, child characteristics such as developmental level and receptive language may also influence these interactions. Sixty-one children entered their respective interventions developmentally 24 months of age or below (Iacono et al., 1998; Nunes & Hanline, 2007; Romski et al., 2010). Overall, these participants increased from zero instances of AAC use or vocalizations at study entry to 5 to 14 spoken words or words produced using AAC during treatment. These studies indicate that preschool children with complex communication needs who are minimally verbal and developmentally 2 years of age or younger can make gains in communication through brief targeted interventions. Children who are minimally verbal (i.e., less than 20 functional spontaneous expressive symbols; Tager-Flusberg & Kasari, 2013) are a unique population of children who are underrepresented in the literature. Little is known about the development of language for school-age children who remain minimally verbal. Notably, preschool children who had some spoken language were the majority of participants. For these children, AAC can be viewed as a way to jump start communicative development. Missing from these studies are older individuals using AAC who continue to have minimal communicative skills and who are often viewed as lifetime users of AAC. It is possible that these dyads may require a different dosage or type of intervention to facilitate gains.

Furthermore, the development of receptive language skills may be essential for this specific population. Increasing receptive language may help children navigate the demands of their daily world, even if language production remains limited. Therefore, receptive language outcomes may be particularly important targets for children who are minimally verbal and using AAC. Yet, it is notable that no study measured children’s receptive language post intervention. Therefore, we have yet to understand how current partner training programs are influencing children’s receptive language. The inclusion of receptive language measures in future studies will help demonstrate the efficacy of current practices on children’s understanding of language.

Strengths and Limitations

Although a variety of adult partners were included, a limitation of this review is the exclusion of studies including peers as communication partners. Due to the unique protocols required to train peers in comparison with adult partners, these studies were excluded. A strength of this review is the inclusion of gray literature. Due to the preliminary state of the literature, this literature was included to provide a comprehensive review. Another strength of the review is the evaluation of methodological quality. Although all published rating scales have limitations, the chosen scales were selected because they included items that represent critical issues in the design and execution of quantitative research. The scales are limited in that they provide equal weight to all items although some items may be considered of greater importance to methodological quality (e.g., experimental control). Yet, this evaluation of quality helps provide context for the interpretation of study findings. The methodological limitations of all studies receiving “weak” quality ratings are important to consider when interpreting findings.

Future Directions

The studies included in this review demonstrate concurrent increases in communication partners’ skills and children’s communication skills. Demonstration of a directional association between the increasing skills of the partner and the children’s increasing communication skills is an important next step. In addition, it will also be important to clearly delineate the instructional strategies provided to partners as well as differential children’s communicative initiations versus responses. Both sets of skills are necessary for successful social interactions; however, it is not clear within the included studies whether children developed both functions. Finally, the strength of the evidence is currently limited by the small number of group design studies, including RCTs. RCTs are a viable design for large-scale trials of existing interventions as well as comparative trials and are necessary to increase the strength of the evidence provided for partner training.

Conclusion

Overall, the effectiveness of interventions focusing on AAC usage will be measured by the interactions of children using AAC with their communication partners in their everyday environment. The current review demonstrates that intervention research designed to support communication partners provides positive preliminary evidence for partners’ adoption of strategies to support children using AAC. Further investigation of partner training programs within larger-scale studies of high methodological quality will help inform and refine intervention protocols. Additional examination of partner training strategies and focus on the context and functions of communication targeted may help guide the merger of these strategies into existing interventions targeting spoken language. Future research may examine how this body of literature and intervention practices may be modified to include children using AAC systems and support their communicative development.

Footnotes

Authors’ Note

Nancy Jones is now at Neuren Pharmaceuticals. The work presented here was completed while at the University of California Los Angeles Department of Applied Linguistics.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The first author received fellowship funds from the Canadian Institutes of Health Research (Doctoral Foreign Study Award) and Autism Speaks (Weatherstone Pre-Doctoral Fellowship) throughout the preparation of this manuscript.

References

Abidin

R. R.

(1995). Parenting Stress Index. Odessa, FL: Psychological Assessment Resources.

American Academy for Cerebral Palsy and Developmental Medicine. (2008). Methodology to develop systematic reviews: Treatment interventions. Retrieved from http://www.aacpdm.org/UserFiles/file/systematic-review-methodology.pdf

Benigno

J. P.

McCarthy

J. W.

(2012). Aided symbol-infused joint engagement. Child Development Perspectives, 6, 181–186.

Beukelman

Mirenda

(2005). Augmentative and alternative communication: Supporting children and adults with complex communication needs (3rd ed.). Baltimore, MD: Brookes.

*Binger

Kent-Walsh

Berens

Del Campo

Rivera

(2008). Teaching Latino parents to support the multi-symbol message productions of their children who require AAC. Augmentative & Alternative Communication, 24, 323–338.

*Binger

Kent-Walsh

Ewing

Taylor

(2010). Teaching educational assistants to facilitate the multisymbol message productions of young students who require augmentative and alternative communication. American Journal of Speech-Language Pathology, 19, 108–120.

*Bingham

M. A.

Spooner

Browder

(2007). Training paraeducators to promote the use of augmentative and alternative communication by students with significant disabilities. Education and Training in Developmental Disabilities, 42, 339–352.

Brady

N. C.

Herynk

J. W.

Fleming

(2010). Communication input matters: Lessons from prelinguistic children learning to use AAC in preschool environments. Early Childhood Services, 4, 141–154.

*Cafiero

J. M.

(1995). Teaching parents of children with autism picture communication symbols as a natural language to decrease levels of family stress (Unpublished doctoral dissertation). The University of Toledo, OH.

10.

Carrow-Woolfolk

(1999). Test for auditory comprehension of language (3rd ed.). Austin, TX: Pro-Ed.

11.

*Chang

G. T. M.

(2009). A visually based naturalistic communication intervention for nonverbal persons with autism (Unpublished doctoral dissertation). Claremont Graduate University, CA.

12.

Cohen

(1992). A power primer. Psychological Bulletin, 112, 155–159.

13.

*Douglas

S. N.

Light

McNaughton

D. B.

(2012). Teaching paraeducators to support the communication of young children with complex communication needs. Topics in Early Childhood Special Education, 33, 91–101.

14.

Fenson

Dale

P. S.

Reznick

J. S.

Thal

Bates

Hartung

J. P.

Reilly

J. S.

(1993). MacArthur Communicative Development Inventories: User’s guide and technical manual. Baltimore, MD: Brookes.

15.

*Ganz

J. B.

Goodwyn

F. D.

Boles

M. M.

Hong

E. R.

Rispoli

M. J.

Lund

E. M.

Kite

(2013). Impacts of a PECS instructional coaching intervention on practitioners and children with autism. Augmentative & Alternative Communication, 29, 210–221.

16.

Granlund

Bjorck-Akesson

Wilder

Ylven

(2008). AAC interventions for children in a family environment: Implementing evidence in practice. Augmentative & Alternative Communication, 24, 207–219.

17.

Higgins

J. P. T.

Green

(2011). Cochrane handbook for systematic reviews of interventions (Version 5.1.0). The Cochrane Collaboration. Available from www.cochrane-handbook.org

18.

Howlin

Gordon

R. K.

Pasco

Wade

Charman

(2007). The effectiveness of Picture Exchange Communication System (PECS) training for teachers of children with autism: A pragmatic group randomized controlled trial. Journal of Child Psychology and Psychiatry, 48, 472–481.

19.

*Iacono

T. A.

Chan

J. B.

Waring

R. E.

(1998). Efficacy of a parent-implemented early language intervention based on collaborative consultation. International Journal of Language & Communication Disorders, 33, 281–303.

20.

Jadad

A. R.

Moore

Carroll

Jenkinson

Reynolds

J. M.

Gavaghan

D. J.

McQuay

H. J.

(1996). Assessing the quality of reports of randomized clinical trials: Is blinding necessary? Controlled Clinical Trials, 17, 1–12.

21.

Kasari

Kaiser

Goods

Nietfeld

Mathy

Landa

. . . Almirall

(2014). Communication interventions for minimally verbal children with autism: Sequential multiple assignment randomized trial. Journal of the American Academy of Child and Adolescent Psychiatry, 53, 635–646.

22.

*Kent-Walsh

Binger

Hasham

(2010). Effects of parent instruction on the symbolic communication of children using augmentative and alternative communication during storybook reading. American Journal of Speech-Language Pathology, 19, 97–107.

23.

Kent-Walsh

McNaughton

(2005). Communication partner instruction in AAC: Present practices and future directions. Augmentative & Alternative Communication, 21, 195–204.

24.

Koppenhaver

D. A.

Erickson

K. A.

Skotko

B. G.

(2010). Supporting communication of girls with Rett syndrome and their mothers in storybook reading. International Journal of Disability, Development and Education, 48, 395–410.

25.

Light

Drager

(2007). AAC technologies for young children with complex communication needs: State of the science and future research directions. Augmentative & Alternative Communication, 23, 204–216.

26.

Logan

L. R.

Hickman

R. R.

Harris

S. R.

Heriza

C. B.

(2008). Single-subject research design: Recommendations for levels of evidence and quality rating. Developmental Medicine & Child Neurology, 50, 99–103.

27.

Mirenda

(2003). Toward functional augmentative and alternative communication for students with autism: Manual signs, graphic symbols, and voice output communication aids. Language, Speech, and Hearing Services in Schools, 34, 203–216.

28.

*Nunes

Hanline

M. F.

(2007). Enhancing alternative and augmentative communication use of a child with autism through a parent-implemented naturalistic intervention. International Journal of Disability, Development and Education, 54, 177–197.

29.

Ospina

M. B.

Seida

J. K.

Clark

Karkhaneh

Hartling

Tjosvold

. . . Smith

. (2008). Behavioural and developmental interventions for autism spectrum disorder: A systematic review. PLoS ONE, 3, 1–32.

30.

Parker

R. I.

Vannest

K. J.

Brown

(2009). The improvement rate difference for single-case research. Exceptional Children, 75, 135–150.

31.

Patterson

S. Y.

Smith

Mirenda

(2012). Parent training programs for families of children with autism: A systematic review of single case contributions. Autism: International Journal of Research and Practice, 16, 498–522.

32.

Roberts

M. Y.

Kaiser

A. P.

(2011). The effectiveness of parent-implemented language interventions: A meta-analysis. American Journal of Speech-Language Pathology, 20, 180–199.

33.

Romski

Sevcik

R. A.

Adamson

L. B.

Cheslock

Smith

(2006). Parents can implement AAC interventions: Ratings of treatment implementation across early language intervention. Early Childhood Services, 1, 249–259.

34.

*Romski

Sevcik

R. A.

Adamson

L. B.

Cheslock

Smith

Barker

R. M.

Bakeman

(2010). Randomized comparison of augmented and nonaugmented language interventions for toddlers with developmental delays and their parents. Journal of Speech, Language, and Hearing Research, 53, 350–364.

35.

*Rosa-Lugo

L. I.

Kent-Walsh

(2008). Effects of parent instruction on communicative turns of Latino children using augmentative and alternative communication during storybook reading. Communication Disorders Quarterly, 30, 49–61.

36.

Smith

Scahill

Dawson

Guthrie

Lord

Odom

. . . Wagner

(2007). Designing research studies on psychosocial intervention in autism. Journal of Autism and Developmental Disorders, 37, 354–366.

37.

Snell

M. E.

Chen

Hoover

(2006). Teaching augmentative and alternative communication to students with severe disabilities: A review of intervention research 1997-2003. Research and Practice for Persons with Severe Disabilities, 31, 203–214.

38.

*Stiebel

(1999). Promoting augmentative communication during daily routines: A parent problem-solving intervention. Journal of Positive Behavior Intervention, 1, 159–169.

39.

Tager-Flusberg

Kasari

(2013). Minimally verbal school-age children with autism: The neglected end of the spectrum. Autism Research, 6, 468–478.

40.

Tomasello

(2001). Perceiving intention and learning words in the second year of life. In Bowerman

Levinson

(Eds.), Language acquisition and conceptual development (pp. 133–158). New York, NY: Cambridge University Press.