Speech-Language Pathologists’ Practices in Augmentative and Alternative Communication during Early Intervention

Abstract

This survey study examined augmentative and alternative communication (AAC) practices reported by early intervention speech-language pathologists (SLPs) across the United States (N = 376). The study examined (a) types of AAC that SLPs reported using (i.e., sign language, photographs, pictures, symbols, talking switches, and iPad apps or dedicated speech-generating devices); (b) SLPs’ perspectives on the influence of child spoken language ability on AAC recommendations; (c) factors that influenced AAC decision-making within early intervention; and (d) perceived barriers associated with AAC implementation. SLPs reported that they were significantly more likely to introduce all types of AAC to children without spoken language abilities compared to children in later stages of language development. On average, they were most likely to report using or recommending sign language and photographs, and least likely to report using or recommending talking switches or speech-generating devices. Of the options provided, child expressive and receptive language abilities were rated as the most important factors to consider when determining AAC use, followed by cognitive ability, diagnosis, and chronological age. SLPs identified caregiver buy-in and carryover across providers as the most significant barriers to AAC implementation. Recommendations for future research and current AAC practices within early intervention are discussed.

Augmentative and alternative communication (AAC) can augment (i.e., supplement) speech or provide an alternative method for communication by replacing speech, and serves as a means to improve communicative success for those with speech intelligibility and/or language impairments (American Speech-Language-Hearing Association, n.d.). In the United States, Part C (2004) of the Individuals with Disabilities Education Act (IDEA) mandates publicly funded early intervention services for children with disabilities under 3 years-old and requires that, when necessary, these services include assistive technology, including AAC. An American Speech-Language-Hearing Association (ASHA) Leader article challenged speech-language pathologists (SLPs) to introduce AAC earlier, including in the birth to age 3 years-old period (Davidoff, 2017). This challenge reflects research on promising AAC interventions for young children (Romski et al., 2010; Solomon-Rice & Soto, 2014). Despite the recent increase in attention to this topic, specific AAC systems used in early intervention, as well as the factors that may influence an SLP’s decision to use AAC, remain largely unknown. While a number of professionals may contribute to the success of AAC interventions, it is particularly vital to study SLPs’ perspectives and practices, given their central role in identifying individuals who may benefit from AAC, conducting comprehensive assessments, and implementing interventions that maximize effective communication (ASHA, n.d.). To address gaps in our understanding of clinical practice and identify future research directions to improve AAC success, we must first begin to identify the types of AAC SLPs report using during early intervention, as well as potential barriers to AAC interventions.

There is a range of AAC options within early intervention. AAC can be unaided, such as sign language or gestures; or aided, such as pictures, photographs, graphic symbols, talking switches (e.g., single- and multiple- message recordable/digitized devices), speech-generating devices (SGDs), or apps on iPads ¹/tablets. There are pros and cons associated with each and children with language impairments can benefit from a range or combination of AAC options (Morin et al., 2018). The research, availability, cost, and training required for AAC vary, making the decision to use AAC complex. Research has shown language gains via sign language (Lederer & Battaglia, 2015; Wright et al., 2013) and graphic symbols (Lerna et al., 2014; Yoder & Stone, 2006). There is also research showing benefits from talking switches (Schweigert & Rowland, 1992) and SGDs (Lorah, 2016; Romski et al., 2010). It is not clear whether one type of AAC takes precedence over another within early intervention. Understanding SLP perspectives and barriers to AAC use (e.g., cost, access) will provide information from which future work can build on, such as by testing the efficacy of specific AAC systems within interventions and addressing identified barriers.

In the United States, early intervention services are offered to any eligible child through publicly funded services at either a free or reduced cost and are the precursor to school-based services (Centers for Disease Control and Prevention, 2019). Of note, children are also able to receive private EI services (e.g., clinics). Historically, children in early intervention have been considered to be too young to benefit from or even be introduced to AAC (Cress & Marvin, 2003). That notion has changed and, in recent years, there has been a push for early intervention to include AAC (Davidoff, 2017; Romski et al., 2010, 2015). A number of studies have found that AAC can support language development in young children ages 1 to 3 years (Branson & Demchak, 2009; Romski et al., 2010; Solomon-Rice & Soto, 2014). Previous work across AAC systems has demonstrated efficacy in increasing communicative turns and expressive language in children under 3 years-old (Romski et al., 2010; Solomon-Rice & Soto, 2014; Wright et al., 2013; Yoder & Stone, 2006). AAC can also support receptive and expressive language growth and promote word combinations and grammar development for preschool- and school-age children with language impairments resulting from various etiologies (Binger & Light, 2007; Brady, 2000; Drager et al., 2006; Kasari et al., 2014). Some AAC systems naturally provide a visual modality (i.e., pictures or graphic symbols), which can aid in language comprehension. Furthermore, AAC can provide users with speech difficulties (e.g., apraxia of speech, dysarthria, phonological impairment) with a modality to communicate more effectively or clarify their spoken language, and provide an opportunity for users to combine symbols and develop grammatical structures.

Research-to-date suggests that children under 3 years-old benefit from AAC interventions, including those that incorporate parent training (Romski et al., 2010, 2015). Early intervention programs often utilize a parent-coaching model, which is consistent with current research demonstrating the positive effects of parent-coached interventions on child language abilities (Roberts & Kaiser, 2011, 2015). Parent coaching allows for language strategies to be incorporated into routines by caregivers. AAC interventions that include caregiver training have shown promise in terms of caregiver ability to implement AAC strategies and child expressive language gains (Kent-Walsh et al., 2010, 2015; Romski et al., 2010). However, research on AAC practices in early intervention is limited and has focused on assistive technology (i.e., assistive and adaptive equipment including but not limited to AAC) across developmental areas through broadly surveying providers, including physical therapists, occupational therapists, SLPs, and teachers (Dugan et al., 2006; Wilcox et al., 2006). Dugan et al. included one survey item on provider decision-making in the following scenario: “A child has normal hearing, babbles rarely, struggles to vocalize, and has not yet produced any speech” (p. 28). Providers were given a limited selection of intervention options ranging from targeting child vocalizations to introducing technology-based AAC and were asked to select which approach they would utilize for a child at 12–24 months and 24–36 months. For children aged 12–24 months, early intervention providers focused primarily on skill development (e.g., vocalizations or sign language), with limited use of aided AAC. For children ages 24–36 months who were not yet speaking, the majority of providers focused on AAC that included pictures or talking switches followed by skill development and then SGDs. Dugan et al. did not investigate SLP practices separately from other providers, which is important because different providers have distinct roles and responsibilities in relation to AAC interventions (ASHA, n.d.). Nevertheless, their findings indicated that these providers tended to focus on spoken language strategies or sign language for children under 24 months of age and implemented aided AAC when children were between 24–36 months and not yet speaking.

In a separate survey of early intervention providers across disciplines, Wilcox et al. (2006) examined assistive technology use (including but not limited to AAC). They found that the majority of SLPs felt there were children on their caseload who should have been using assistive technology but were not currently doing so. SLPs were more likely than other providers (e.g., physical therapists, teachers) to agree with a statement that young children needed certain skills before using assistive technology. Notably, across disciplines, providers with more training in assistive technology were less likely to agree that young children needed certain prerequisite skills as compared to providers with less training; thus, it appears there is some divergence in provider beliefs, but it is not yet clear whether these same patterns hold when surveying SLPs and examining AAC exclusively, in addition to what factors may influence beliefs.

While not specific to early intervention or SLP perspectives, two surveys by Webb et al. (2019a, b) examined AAC professional decision-making surrounding AAC use. While several child characteristics and AAC attributes were identified, language, cognitive, learning abilities, and support for AAC from communication partners were more important considerations compared to physical abilities and AAC attributes (e.g., durability, cost). Using vignettes, Webb et al. (2019b) found that AAC choices were influenced more by child motivation to communicate via AAC and anticipated progress than language ability. Although these studies did not focus on early intervention, their findings represent the complex decision-making process and range of factors impacting a professional’s decision regarding AAC use.

SLPs face barriers associated with implementing AAC (Donato et al., 2018). Caregivers may be reticent to trial AAC due to concerns that AAC will hinder spoken language development, although this is not supported by research (Binger et al., 2008; Kasari et al., 2014; Romski et al., 2010). Training poses another potential barrier, as the training needed to implement AAC can be extensive for SLPs and caregivers. Most SLPs are given limited time for professional development (O’Connor & Pettigrew, 2009), and may not be able to prioritize rigorous AAC training. Because a minority of children will go on to be minimally speaking or have low speech intelligibility requiring AAC as their primary mode of communication (ASHA, 2018, schools survey), SLPs may not devote their limited time to AAC-focused professional development. Yet, a 2017 survey of early intervention SLPs found that 29% of respondents identified AAC as a preferred area for professional development (ASHA, 2018, minisurvey). In addition, SLPs may not have access to necessary resources. In the United States, early intervention SLPs often serve children through publicly funded programs, which may not have funding for a variety of AAC systems. While AAC is covered through some insurance plans in the United States, coverage is not universal, and SLPs must provide extensive documentation that the child demonstrates AAC proficiency prior to insurance approval, which is not possible without appropriate access to AAC devices (ASHA, n.d.). It is important to determine which barriers early intervention SLPs identify as most problematic to AAC implementation.

In summary, AAC can be effective for supporting language development in early intervention but findings to date indicate limited use of AAC for children in this age range (Dugan et al., 2006). In addition, it is unclear what specific AAC options SLPs recommend and utilize in early intervention, what factors influence their decision-making, and what potential barriers to AAC use exist. While evidence-based practices in the field of speech-language pathology include consideration of current research, clinical expertise, and family values, advancement of these practices and the development of studies that empirically test current AAC practices and address barriers depends on first understanding what SLPs are doing and why they are making certain decisions with regard to AAC and early intervention. To begin addressing these gaps in research, this study probed SLPs’ perspectives and reported practices. The following research questions were asked:

1. Do SLPs recommend certain AAC options over others and do these recommendations differ on the basis of the child’s expressive language stage, including (a) not yet speaking, (b) speaking in single words, and (c) speaking in two-word utterances? Specifically, this question probed whether SLPs reported recommending certain AAC options more frequently than others, as well as how these reported recommendations changed based on the child’s expressive language ability.

2. What child characteristics most strongly influence the use of AAC among SLPs? Specifically, what characteristics did SLPs report as most strongly influencing their decisions surrounding AAC?

3. What factors do SLPs identify as influencing their use of AAC within interventions? Specifically, what child factors do SLPs identify as influencing AAC implementation and what barriers do SLPs identify to AAC implementation?

Method

Participants

SLPs were recruited to participate in an online survey via social media postings in early intervention and SLP groups, as well as by word-of-mouth. The first page of the online survey included a consent form where participants were informed that the study was voluntary and they could stop at any time. Participants could only continue to the survey if they consented to participate in the study. Inclusionary criteria were: (a) currently living within the United States of America, and (b) currently a practicing SLP working with at least one child between the ages of birth to 3 years-old.

The survey was focused on SLPs rather than all early intervention practitioners for the following reasons: (a) previous research has more broadly surveyed the latter providers (e.g., Dugan et al., 2006); and (b) the researchers were particularly interested in how decision-making regarding AAC use was related to child language ability, which falls under an SLP’s scope of practice (ASHA, n.d.).

A total of 376 SLPs participated in the survey, representing 47 states. They were an average of 35.77 years of age (SD = 10.36, range: 30–65), had an average of 9.66 years (SD = 9.49, range: <1 – 42) of experience as an SLP, and an average of 6.83 years (SD = 7.13, range: <1 – 42) of experience as a provider to children from birth to 3 years. The majority of SLPs utilized parent coaching (84.8%) and worked in a government-funded setting (66.1%). Data on participant gender were not collected. See Table 1 for additional participant descriptives.

Table 1.

Participant Descriptives (n = 376)

Variable	Value	Range
Chronological age (years)	M = 35.77, SD = 10.36	30–65
Years of SLP experience	M = 9.66, SD = 9.49	<1 – 42
Years of experience as birth to three provider	M = 6.83, SD = 7.13	<1 – 42
Average hours/week providing birth to three services	M = 23.29, SD = 13.64	.50 – 50
Average caseload size	M = 17.32, SD = 13.26	1 – 100
Percentage of SLPs utilizing parent coaching	84.8%	-
Work setting (% government-funded)	66.1%	-

Research Design

This study utilized survey methodology for the following reasons: Given the relatively limited information available on this topic, an online survey offered fewer constraints related to scheduling participant visits and made it possible to sample a relatively large number of SLPs from geographically diverse areas within the United States.

The study was deemed exempt by the University of Wisconsin – Madison Institutional Review Board because no identifying information was collected.

Materials

Survey

The online survey consisted of 29 questions pertaining to language input strategies and recommendations as well as AAC use and practices within early intervention (defined as intervention for children from birth to 3 years-old). The survey was administered via Qualtrics (an online platform) and the anticipated completion time was 15 min (Qualtrics, 2017). The questions regarding language input strategies and recommendations were adapted from previous work (Venker et al., 2019, 2020) and the results are reported elsewhere (Maltman et al., under review). The study utilized multiple choice and open-ended questions. Participants could skip any questions they did not want to answer. Please see the supplemental materials for a copy of the survey questions included in this study.

Participant Characteristics.

The survey included 10 questions to obtain demographic and background information, including participant age, years of experience as an SLP, years of experience as a provider to children aged birth to 3 years-old, whether they used parent coaching (yes/no), whether they worked in a government-funded agency, and their hours per week and caseload size within the birth to 3 population.

AAC Recommendations by Child Language Level.

SLPs were asked to rate their likelihood of recommending six different AAC tools for children at three stages of expressive language: (a) not yet saying words consistently, (b) primarily producing single words, and (c) primarily producing two-word utterances. Specifically, SLPs were asked a set of three questions: “For a child [not yet saying words consistently OR primarily producing single words to communicate OR primarily producing two-word utterances to communicate], how likely are you to use or recommend that caregivers use the following tools to support the child’s communication development: (a) sign language, (b) photographs, (c) pictures from the internet or magazines, (d) symbols from software programs used for creating communication boards (e.g., Boardmaker²), (e) basic switches, such as a BIGmack³ or Step-by-Step⁴, and (f) iPad application or dedicated speech-generating device?” All questions were analyzed on a five-point scale: 1 (extremely unlikely), 2 (somewhat unlikely), 3 (neither likely nor unlikely), 4 (somewhat likely), and 5 (extremely likely), where higher numbers indicated a greater likelihood of recommending a specific AAC system. A total of 334 SLPs completed the ratings regarding children not yet saying words consistently, 273 SLPs completed the ratings for children primarily producing single words to communicate, and 230 SLPs completed the ratings for primarily producing two-word utterances to communicate. Of note, the exclusion of “tablet” in the “iPad application or dedicated speech-generating device” response was unintentional in the survey design.

Child Characteristics and AAC Use.

SLPs were asked to rate the importance of five child characteristics (i.e., expressive language ability, receptive language ability, cognitive ability or mental age, diagnosis, and chronological age) in their decisions regarding AAC use. Specifically, they were asked: “When determining whether or not to use aided or unaided augmentative and alternative communication (AAC) with a child, in your opinion, how important are the following child characteristics?” Each child characteristic was listed separately and were rated as follows: 1 (not important), 2 (slightly important), 3 (moderately important), and 4 (very important), and 205 SLPs completed all ratings for this question.

Open-ended Questions.

The survey included two open-ended items inquiring about the decision to use any AAC and barriers associated with AAC interventions. The first item read: “Please explain what specific child characteristics result in your decision to use AAC with a child (e.g., specific diagnoses, age, language abilities, etc.).” SLPs typed their responses into a text box; 124 responded. The second question asked: “What is the primary barrier associated with AAC interventions?” SLPs again were able to type their responses in a text box, and 128 of the SLPs provided a response. Notably, they could report more than one child characteristic influencing their decision or more than one barrier associated with AAC interventions. For example, a participant could state that both child language ability and motor skills influenced their decision. For these questions, we were interested in SLPs’ reports when provided with open-ended questions in order to identify other factors that may not have been listed in Research Question 2 as well as the barriers SLPs independently identified.

Procedures

Data Collection and Analysis

Nonparametric statistics were utilized because the data were categorical. To reduce Type I Error, Holm-Bonferroni corrections were applied and corrected p-values for each set of analyses were reported (Holm, 1979). Nonparametric effect size r was also reported to aid in interpretation of the results; r was interpreted as follows: r =.10 small effect, r =.30 moderate effect, and r = .50 large effect (Pallant, 2007).

Research Question 1: AAC Recommendations by Child Language Stage.

Descriptive statistics were used to indicate the percentage of participants who selected each response [from 1 (extremely unlikely) to 5 (extremely likely)] regarding their likelihood of recommending each AAC tool for children within the three expressive language stages. To determine if SLPs were more likely to use or recommended certain types of AAC for children within each language stage (i.e., Did SLPs report that they were more likely to use or recommend sign language than talking switches for children producing single word utterances?), Wilcoxon signed-rank tests were used to compare all AAC recommendations separately for each language level. For the second part of Research Question 1 (i.e., if SLPs reported they were more likely to recommend specific AAC tools based on child language level), Friedman’s ANOVAs were conducted across the three language levels separately for each AAC tool. Friedman’s ANOVAs were used to account for each participant rating their likelihood of recommending a specific AAC option under three separate conditions (i.e., language levels; Riffenburgh, 2006). Significant results were followed up with Wilcoxon signed-rank tests to compare specific language levels.

Research Question 2: Child Characteristics and AAC Use.

Descriptive statistics were used to indicate the percentage of SLPs who selected each category [from 1 (not important) to 4 (very important)] regarding child expressive language ability, receptive language ability, cognitive ability or mental age, diagnosis, and chronological age. Next, Wilcoxon signed-rank tests were used to compare ratings between each characteristic to determine which child characteristics SLPs rated as most-to-least important in influencing any AAC use.

Research Question 3: Child Factors and Barriers to AAC Interventions.

SLPs were asked what child factors influenced their decision to use AAC. Because this question was open-ended, it was necessary to classify responses into categories in order to quantitatively describe the responses. The responses were reviewed separately by the first and second authors; the second author identified eight categories and the first author identified nine categories (i.e., the same eight categories plus an additional category; 88.89% agreement on initial category conceptualization). Specifically, the two authors separately identified eight conceptually consistent categories, which were subsequently agreed upon through discussion and named. The ninth category identified by the first author (Chronological Age) was conceptually distinct from the other eight identified categories, and after discussion, was added to the list. This process resulted in responses being categorized into nine categories that captured the majority (96.8%) of the responses. Categories included: Receptive and/or Expressive Language, Diagnosis and/or Prognosis, Cognitive Abilities, Motor Skills, Environmental Supports, Chronological Age, Limited Success with Other Approaches, Frustration, and AAC Benefits Any Child. The percentage of responses for each category was calculated. When a respondent indicated multiple child factors, each factor was considered separately.

A similar approach was used for the second part of Research Question 3, which asked what SLPs believed to be the primary barrier to AAC interventions. Once again, responses were reviewed by the first and second authors separately and then a discussion ensued during which common themes (i.e., conceptually consistent categories) were identified and subsequently named. This resulted in the following categories: Caregiver Buy-In, Carryover Across Providers, Cost, Access, Child Factors, SLP and/or Parent Training, and Time (100% agreement on initial category conceptualization). These categories captured the majority of responses (95.3%). The percentage of responses that fit into each category out of the total number of responses were calculated. For each open-ended question, a sample of direct quotations of SLP responses were provided to enrich the descriptive information. The first and second authors reviewed all open-ended responses and the presented quotations represent the broad range of SLPs’ responses.

Results

Research Question 1: AAC Recommendations by Child Language Stage

SLPs were asked to rate how likely they would be to use or recommend six AAC options for children at three expressive language stages (see Table 2 and Figure 1 for descriptive data across AAC types and language levels). Similar but slightly different patterns emerged (see Table 3). For children not yet using words, SLPs were significantly most likely to report recommending sign language followed by photographs, symbols from communication boards, pictures from the internet or magazines, iPad apps or speech-generating devices, and talking switches. For children producing single words, SLPs were significantly most likely to report recommending sign language or photographs followed by symbols from communications boards or pictures from the internet or magazines, iPad apps or speech-generating devices, and talking switches. For children producing two-word utterances, SLPs were significantly most likely to report recommending photographs followed by sign language or pictures from the internet or magazines, symbols from communication boards, iPad apps or speech-generating devices, and talking switches.

Table 2.

Percentage of Recommended AAC Systems by Child Language Level, From 1 (extremely unlikely) to 5 (extremely likely)

AAC system	1	2	3	4	5
Not yet saying words consistently
Sign language (M = 4.51, SD = 0.81)	0.6%	3.6%	5.7%	24.7%	65.5%
Photographs (M = 4.22, SD = 0.84)	0.6%	4.5%	9.5%	43.5%	42.0%
Pictures from internet/magazine (M = 3.30, SD = 1.20)	7.7%	20.2%	23.2%	31.5%	17.3%
Symbols from communication boards (M = 3.48, SD = 1.15)	6.9%	15.8%	17.3%	42.7%	17.3%
Basic (talking) switches (M = 2.61, SD = 1.09)	16.2%	33.5%	26.9%	19.5%	3.9%
iPad app or speech-generating device (M = 2.78, SD = 1.17)	14.3%	33.0%	19.3%	27.1%	6.3%
Producing single words
Sign language (M = 3.81, SD = 1.17)	3.6%	15.6%	10.9%	35.6%	34.2%
Photographs (M = 3.71, SD = 1.13)	4.0%	15.2%	12.7%	41.7%	26.4%
Pictures from internet/magazine (M = 3.03, SD = 1.30)	14.7%	24.5%	19.0%	27.1%	14.7%
Symbols from communication boards (M = 2.96, SD = 1.30)	16.7%	23.9%	19.6%	26.8%	13.0%
Basic (talking) switches (M = 2.05, SD = 1.15)	43.3%	24.9%	17.7%	11.2%	2.9%
iPad app or speech-generating device (M = 2.32, SD = 1.28)	34.9%	26.5%	16.4%	15.6%	6.5%
Producing two-word utterances
Sign language (M = 2.67, SD = 1.29)	20.6%	36.1%	6.4%	30.0%	6.9%
Photographs (M = 2.91, SD = 1.33)	17.7%	30.2%	5.6%	36.6%	9.9%
Pictures from internet/magazine (M = 2.55, SD = 1.28)	25.7%	32.2%	9.1%	27.8%	5.2%
Symbols from communication boards (M = 2.18, SD = 1.15)	34.5%	35.8%	9.5%	18.1%	2.2%
Basic (talking) switches (M = 1.53, SD = 0.81)	62.6%	26.1%	7.4%	3.5%	0.4%
iPad app or speech-generating device (M = 1.79, SD = 1.05)	52.2%	29.7%	6.5%	9.9%	1.7%

Note. Due to rounding, some columns do not add up to 100.0%.

Figure 1.

Proportion of Recommended AAC Systems by Child Expressive Language Level

Table 3.

Comparisons of AAC Recommendations by Child Language Level

	(a) Sign language	(b) Photographs	(c) Pictures from internet/magazine	(d) Symbols from communication boards	(e) Basic (talking) switches
Type of AAC system	z, r	z, r	z, r	z, r	z, r
Not yet saying words consistently
(a) Sign language
(b) Photographs	−5.12***, .28
(c) Pictures from internet/magazine	−11.91***, .65	−11.75***, .64
(d) Symbols from communication boards	−11.03***, .60	−9.29***, .51	2.02*, .11
(e) Basic (talking) switches	−14.65***, .80	−14.05***, .77	−7.80***, .43	−10.02***, .55
(f) iPad app or speech-generating device	−13.89***, .76	−13.18***, .72	−5.87***, .32	−8.55***, .47	2.77*, .15
Producing single words
(a) Sign language
(b) Photographs	−1.40, .08
(c) Pictures from internet/magazine	−7.91***, .48	−9.07***, .55
(d) Symbols from communication boards	−8.46***, .51	−9.02***, .54	−0.84, .05
(e) Basic (talking) switches	−12.70***, .77	−12.82***, .77	−9.66***, .58	−10.18***, .61
(f) iPad app or speech-generating device	−11.51***, .70	−11.66***, .70	−6.86***, .42	−7.91***, .48	4.53***, .27
Producing two-word utterances
(a) Sign language
(b) Photographs	2.98**, .20
(c) Pictures from internet/magazine	−1.28, .08	−5.96***, .39
(d) Symbols from communication boards	−5.03***, .33	−7.24***, .48	−4.31***, .28
(e) Basic (talking) switches	−9.73***, .64	−10.81***, .71	−9.65***, .64	−8.40***, .55
(f) iPad app or speech-generating device	−7.57***, .50	−8.92***, .59	−6.76***, .45	−5.52***, .36	4.55***, .30

Note.

^*p < .050

^**p < .010

^***p < .001 based on Holm-Bonferroni adjusted p-values (Holm, 1979).

Results of Wilcoxon signed-rank tests and nonparametric effect size r reported (Pallant, 2007).

Recommendations for AAC Use Across Language Stages

When comparing AAC recommendations by child language level, SLPs were significantly more likely to report recommending all types of AAC for children with lower expressive language abilities compared to children with more advanced expressive language abilities (see Table 4). Follow-up Wilcoxon signed-rank tests indicated that SLPs were significantly more likely to report recommending all AAC for children not yet using words compared to children producing single and two-word utterances. Similarly, SLPs were significantly more likely to report recommending all AAC for children producing single words compared to children producing two-word utterances.

Table 4.

Recommended AAC System by Child Language Level

	(a) Not yet saying words consistently	(b) Producing single words	(c) Producing two-word utterances	χ 2 F	p	Wilcoxon Signed-Rank tests	Effect size (r)
AAC system	(Mean rank)	(Mean rank)	(Mean rank)
Sign language	2.66	2.07	1.27	300.43	<.001	a > b > c	.57, .80, .69
Photographs	2.48	2.08	1.44	191.10	<.001	a > b > c	.44, .70, .61
Pictures from internet/magazine	2.31	2.03	1.66	85.30	<.001	a > b > c	.30, .49, .36
Symbols from communication boards	2.52	2.06	1.42	209.05	<.001	a > b > c	.47, .71, .59
Basic (talking) switches	2.50	1.97	1.53	177.66	<.001	a > b > c	.48, .69, .46
iPad app or speech-generating device	2.46	1.97	1.57	154.48	<.001	a > b > c	.42, .65, .53

Note. Results presented in the table are based on Holm-Bonferroni adjusted p-values (Holm, 1979). Nonparametric effect sizes corresponding to Wilcoxon signed-rank tests: a vs. b, a vs. c, b vs. c (Pallant, 2007).

Research Question 2: Child Characteristics and AAC Use

Question 2 focused on the child characteristics that SLPs reported as most strongly influencing the use of AAC. Of the survey options presented, the majority of SLPs rated expressive language ability (70.7%), receptive language ability (63.5%), and cognitive ability or mental age (54.1%) as very important (see Table 5). Conversely, only 21.4% of SLPs rated child diagnosis as very important, and 15.6% of SLPs rated chronological age as very important. There was not a significant difference between SLP’s ratings of expressive versus receptive language, z = −1.31, p = .190, r = .09. SLPs rated expressive language ability as significantly more important than cognitive ability or mental age, z = −3.89, p < .001, r = .27, diagnosis, z = −9.14, p < .001, r = .64, or chronological age, z = −10.45, p < .001, r = .73. SLPs rated receptive language ability as significantly more important than cognitive ability or mental age, z = −3.96, p < .001, r = .28, diagnosis, z = −8.75, p < .001, r = .61, or chronological age, z = −10.37, p < .001, r = .72. SLPs rated child cognitive ability or mental age as significantly more important than child diagnosis, z = −7.30, p < .001, r = .51, or chronological age, z = −9.67, p < .001, r = .68. Lastly, SLPs rated child diagnosis as significantly more important than chronological age, z = −4.96, p < .001, r = .35. In summary, SLPs reported viewing expressive and receptive language abilities as the most important child factors to consider when determining any AAC use, followed by cognitive ability, child diagnosis, and chronological age, respectively.

Table 5.

Importance of Child Characteristics in Determining AAC Use

Child characteristic	M (SD)	Range, Median	1 (not important)	2 (slightly important)	3 (moderately important)	4 (very important)
Expressive language ability	3.57 (0.74)	1–4, 4	1.4%	10.6%	17.3%	70.7%
Receptive language ability	3.50 (0.76)	1–4, 4	1.9%	10.1%	24.5%	63.5%
Cognitive ability or mental age	3.31 (0.88)	1–4, 4	4.3%	14.5%	27.1%	54.1%
Diagnosis	2.69 (0.93)	1–4, 3	10.7%	30.6%	37.4%	21.4%
Chronological age	2.37 (0.96)	1–4, 2	18.5%	41.5%	24.4%	15.6%

Research Question 3: Child Factors and Barriers to AAC Interventions

Child Factors

SLPs were asked two open-ended questions, the first being what child factors influenced their decision to use AAC interventions (see Tables 6 and 7). They could provide multiple responses, and their answers were organized into nine categories. Approximately 43% of SLPs reported Receptive and/or Expressive Language, 39% reported Diagnosis and/or Prognosis, 36% reported Cognitive Factors, 21% reported Motor Skills, and 18% reported Environmental Supports. Approximately 16% of SLPs reported Chronological Age and 14% reported Limited Success with Other Approaches. The fewest number of SLPs reported Frustration and AAC Benefits any Child, at 10% and 6%, respectively.

Table 6.

Child Factors and Barriers Reported by SLPs as Influencing AAC Interventions

Categories	n	%
Child factors
Receptive and/or expressive language	53	42.7
Diagnosis and/or prognosis	48	38.7
Cognitive factors	45	36.3
Motor skills	26	21.0
Environmental supports	22	17.7
Chronological age	20	16.1
Limited success with other approaches	17	13.7
Frustration	13	10.5
AAC benefits any child	8	6.5
Barriers
Caregiver buy-in	50	39.1
Carryover across providers	35	27.3
Cost	32	25.0
Access	24	18.8
Child factors	23	18.0
SLP and/or parent training	18	14.1
Time	9	7.0

Note. SLPs could provide multiple responses to each question; thus, percentages do not add up to 100% but instead represent the percentage of SLPs who reported responses in each category out of the total number of SLPs who responded to the survey item.

Table 7.

Direct Quotations of SLP Responses to Open-Ended Questions

Survey item	Response
Please explain what specific child characteristics result in your decision to use AAC with a child (e.g., specific diagnoses, age, language ability).	“I have learned to presume competence.”
	“Children who have severe motor impairments would be strong candidates. Those who can’t communicate due to medical issues (ventilator, trach). Also, children who don’t seem to be making good progress with intervention and have parents open to working with AAC.”
	“Age is a big factor. I do not recommend AAC devices to early intervention in my setting of work. Diagnosis is also important.”
	“The child needs to understand purpose and function.”
	“I wouldn’t withhold AAC from any language delayed child. It∙s another tool to try with practically any child who could benefit from it.”
	“There are no prerequisites required for a child to use AAC. However, child characteristics such as specific diagnosis, age, language ability, etc. will likely impact my decision on the type of AAC to trial.”
	“If child is two and oral language and signs are not working, I evaluate for AAC.”
What is the primary barrier associated with augmentative and alternative communication (AAC) interventions?	“Helping parents to realize that AAC is a tool/process rather than a quick fix. Getting parents and staff to buy into the process of modeling appropriate language on the device and giving appropriate feedback to the child’s attempts to use AAC.”
	“In my area, I find the largest barrier is misconceptions of stakeholders (EI Officials, payors) about prerequisites for SGDs, often leading to use of less robust no-tech/low-tech systems which do not meet as many of the child’s needs, leading to abandonment.”
	“Parent training, AAC resources, learning to generalize using the AAC out in the community and during their daily lives.”
	“Parents think it will limit expressive abilities down to road. It is often difficult to obtain AAC devices other than pictures or using signs with younger children.”
	“Access to a variety of systems to trial and lack of time in daily schedule to set up/make modifications/learn systems. Also, lack of collaboration across various providers!!”

Barriers

SLPs were also asked what they considered as the primary barrier to AAC interventions. Responses were organized into seven categories. Approximately 39% reported Caregiver Buy-In and 27% reported Carryover Across Providers. Approximately 25% of SLPs reported Cost, 19% reported Access, and 18% reported Child Factors. The fewest number of SLPs reported SLP and/or Parent Training and Time, at 14% and 7%, respectively.

Discussion

This study examined SLPs reported use of AAC in early intervention with children at different language stages, the reported influence of child characteristics in SLP decision-making regarding AAC, and the child factors and barriers SLPs’ identified as impacting their recommendations for AAC interventions. SLPs reported that they were more likely to introduce all types of AAC to children with no spoken language abilities compared to children in later stages of expressive language development. When using AAC, they reported that they were most likely to use sign language, photographs, and pictures and least likely to use talking switches or SGDs. Of the options provided, expressive and receptive language abilities were rated as the most important child factors to consider when determining any AAC use. SLPs identified lack of Caregiver Buy-In and Carryover Across Providers as the most significant barriers.

Research Question 1: AAC Recommendations by Child Language Stage

SLPs were significantly more likely to recommend and/or introduce AAC to children who were not speaking compared to those with spoken language. This finding is consistent with a study that included children with cerebral palsy showing that emerging and established talkers were less likely to have AAC goals compared to children not talking (Smith & Hustad, 2015). Research has focused primarily on AAC interventions for children with little to no intelligible spoken language (Kent-Walsh et al., 2010; Romski et al., 2010; Solomon-Rice & Soto, 2014; Wright et al., 2013). Research regarding AAC interventions in young children producing single- or two-word utterances is not as strong; thus, SLPs may choose to utilize a number of other effective approaches (e.g., language expansions, modeling word combinations) over introducing AAC for these children (Roberts & Kaiser, 2015). Taken together, it may be that when children are beginning to speak, SLPs are more likely to work with parents on strategies that explicitly target spoken language abilities.

Recommendations for Each AAC Option Across Language Stages

SLPs were more likely to report recommending sign language, photographs, or pictures over talking switches and SGDs regardless of child language level. This finding aligns with a previous study where early intervention SLPs were more likely to report using pictures, objects, or graphic symbols over SGDs (Dugan et al., 2006). However, van der Meer et al. (2012) suggests there may be greater gains using SGDs/iPads over other AAC options. In addition, another study found that paraprofessionals and teachers preferred iPads over pictures for school-age children (Lorah, 2016), although these preferences have not been explored in early intervention; the language skills and communication demands of older children versus children differ. The reported use of and/or recommendations for sign language, photographs, and pictures over talking switches or SGDs in early intervention may reflect the fact that these options are more accessible and less costly than SGDs. Sign language, which was the most common type of AAC SLPs reported using or recommending for children with no words or only single words, is readily available and may be more accepted by caregivers, given the mainstream use of baby signs in today’s society (Pizer et al., 2007). Photographs or pictures are also relatively low-cost options and are evidence-based strategies for supporting language development in young children (Romski et al., 2015; Wright et al., 2013). It will be important to continue examining AAC use and preferences across child age and language ability in future research.

Limited training has been posited as a reason for low AAC use in early intervention, particularly regarding SGDs (Davidoff, 2017). SLPs may receive little training in AAC, and only a small percentage of children use/need SGDs throughout the lifespan (ASHA, 2018, schools survey; Binger & Light, 2006). Yet, the percentage of individuals who go on to use AAC in preschool and beyond may not reflect the actual level of need because some children may benefit from AAC but do not receive AAC interventions (Binger & Light, 2006). The findings of the current study, in combination with previous work, suggest that aided AAC is often not considered unless children have no spoken language or show minimal spoken language gains (Dugan et al., 2006; Smith & Hustad, 2015).

Research Question 2: Child Characteristics and AAC Use

SLPs rated expressive and receptive language as the most important factors when considering AAC, a finding that is consistent with a survey of AAC professionals serving individuals across the age range (Webb et al., 2019a). The importance of expressive language is not surprising, given that AAC is often recommended for individuals who have little to no spoken language. The finding that ratings of expressive and receptive language were not significantly different was surprising, although this aligns with the idea that receptive language provides a basis for expressive language development (Fenson et al., 1994). SLPs may also recognize that AAC can support receptive language development (Brady, 2000; Drager et al., 2006). Cognitive abilities and child diagnosis were rated as less important than expressive and receptive abilities; this finding is consistent with previous research showing children with and without intellectual disabilities can benefit from AAC (Kent-Walsh et al., 2010; Romski et al., 2010).

Research Question 3: Child Factors and Barriers to AAC Interventions

Child Factors

Consistent with the ratings in Research Question 2, SLPs most frequently reported Receptive and/or Expressive Language as influencing their decision to use AAC followed by Diagnosis and/or Prognosis; thus, SLPs may be more likely to introduce AAC in early intervention if the child’s diagnosis strongly suggests that they may never develop adequate spoken language skills. The use of specific AAC options may also be more commonly utilized in certain populations. For instance, the Picture Exchange Communication System (PECS) has been studied in children with autism spectrum disorder (Lerna et al., 2014; Yoder & Stone, 2006) and sign language is commonly used with children with Down syndrome (Özçalışkan et al., 2016; Wright et al., 2013). Approximately 16% of SLPs listed Chronological Age as a factor. This is consistent with previous work showing providers primarily focused on skill development (i.e., vocalizing or sign language) for children ages 12 to 24 months and were unlikely to introduce aided AAC (Dugan et al., 2006).

Limited Success with Other Approaches was mentioned 14% of the time. It is not necessary to wait until a child does not progress in spoken language to begin trialing AAC. Targeting spoken language or incorporating AAC should not be an “either/or” decision (for a review, see Romski et al., 2015). SLPs can utilize AAC to simultaneously target language and speech skills through the use of articulatory models, while accessing a range of vocabulary. In fact, AAC interventions typically include a spoken and visual model of language (Kent-Walsh et al., 2010; Romski et al., 2010), while also providing an alternative means for communication. Some SLPs in the current study (6%) indicated there were no prerequisites to AAC use and felt every child could benefit from AAC. This is supported by research (Branson & Demchak, 2009; Romski et al., 2010; Solomon-Rice & Soto, 2014) as well as the National Joint Committee for the Communication Needs of Persons with Severe Disabilities under the Communication Bill of Rights (Brady et al., 2016), which argues that all children with communication needs not met by speech alone should have the opportunity to use AAC regardless of perceived level of developmental ability.

Barriers

Early intervention SLPs are uniquely suited to provide parent-coached AAC interventions that take advantage of a child’s natural environment and routines. Working together alongside parents is an important component of early intervention (Kaiser & Roberts, 2013; Roberts & Kaiser, 2015), and parents play a fundamental role in skill generalization and success, including within AAC interventions (Kent-Walsh et al., 2010, 2015); however, SLPs face barriers that may prevent or deter them from using AAC in early intervention. Caregiver Buy-In was frequently reported as a barrier. Among SLPs who identified this barrier, many also reported caregivers’ misconceptions surrounding AAC and its effects on spoken language (see Table 6). While the best available evidence suggests that AAC supports and does not hinder spoken language development (Kaiser et al., 2014; Romski et al., 2010; Wright et al., 2013), some parents may continue to have concerns. Carryover Across Providers was also commonly reported as a significant barrier to AAC use. AAC will be more successful and less likely to result in abandonment when families partner with interventionists (Angelo et al., 1995; Parette et al., 2000).

Clinical Implications

SLPs identified several barriers to AAC interventions (e.g., Cost, Access, Caregiver Buy-In, Time). While these barriers are unlikely to disappear, it is possible to mitigate their impact. For example, early intervention teams could explore different approaches to AAC training, such as weighing the pros and cons of training all members to acquire intermediate-level AAC knowledge versus designating one SLP on a team to develop AAC expertise. SLPs can also work with AAC vendors and app developers to access more expensive AAC options, such as through device rental programs. Clinics may be able to partner together to create an AAC resource library, allowing for access at lower costs. Addressing Caregiver Buy-In may take a nuanced approach. Many effective parent-to-parent mentoring programs exist within other areas (Singer et al., 1999). It may be useful to connect parents of children who use AAC with parents of children who may benefit from AAC. Furthermore, parent education is particularly important. A worthwhile conversation that focuses on a child’s strengths might involve educating parents on the difference between speech and language; in addition, it may be helpful to explain that for some children, language abilities exceed speech and AAC can provide an alternative modality for communication that allows children to produce language when speech is difficult or absent.

Limitations and Future Directions

While this study has many strengths, there were also limitations. The participants were a convenience sample and may not represent the beliefs of all early intervention SLPs. This study utilized a survey to probe practices through SLPs’ perspectives and did not directly assess or observe SLPs’ practices; future research should use more robust methodologies. Furthermore, the AAC terminology varied within the survey (i.e., AAC tools vs. aided or unaided AAC vs. AAC), and tablets were not included under the AAC option “iPad app or speech-generating device,” all of which may have impacted how participants responded to survey items. This study probed AAC use across populations in early intervention so information specific to children with particular diagnoses was not available and represents an important future direction. Furthermore, this study did not examine what AAC was available to the participants, which may have influenced decision making regarding AAC selection. Many SLPs did not answer the open-ended questions, which may have biased the results regarding barriers and factors influencing AAC options and practices.

The field of AAC is relatively young and considerable research is needed to determine optimal AAC practices in early intervention. Comparing the efficacy of different AAC options for children in early intervention at different stages of spoken language development will provide a foundation for SLPs. Future research should also explore cultural considerations and disparities in AAC access based on demographic factors (e.g., race, socio-economic status), given research in other areas demonstrating differences in diagnostic ages and services provided (McGregor, 2020). This study did not examine AAC use from the perspective of caregivers, which is a necessary next step, given SLP reports of Caregiver Buy-In as a frequent barrier. Carryover Across Providers was also identified as a barrier, and other practitioners within early intervention were not surveyed. Collaborations between SLPs and researchers to develop effective parent education programs and build buy-in from other stakeholders (e.g., parents and other providers) represent critical future directions. It is essential to consider other providers’ perspectives and practices, which may help in developing partnerships with parents and service providers to eliminate barriers and build a team that can holistically support the child in developing functional communication across daily environments.

Conclusion

The SLPs who participated in this study most often reported using or recommending sign language, photographs, or pictures over talking switches or SGDs. This study provides important information on current SLP perspectives regarding AAC use in early intervention; findings suggest the importance of future studies assessing the efficacy of different AAC options for young children with differing diagnoses and language abilities. Caregiver Buy-In and Carryover Across Providers were the most common barriers identified. There is a critical need to develop effective strategies to address barriers to AAC use within early intervention, particularly with regard to incorporating stakeholder involvement to increase AAC success.