Speech Impairments Explain Unique Variance in Adaptive Behavior Skills in Young People With Down Syndrome

Abstract

Background

Down syndrome (DS) is a disorder characterized by impairments in global cognitive abilities and adaptive function. In addition, individuals with DS demonstrate pronounced speech and language deficits. However, little is known about the linguistic correlates of impaired adaptive functioning in DS.

Method

Using the Adaptive Behavior Assessment System–Second Edition and the Children's Communication Checklist–Second Edition (CCC-2), this study investigated the unique variance in adaptive skills accounted for by speech and language impairments in individuals with DS (N = 29, M _age = 13.46).

Results

Pearson correlations revealed that a composite of CCC-2 structural language scales, but not pragmatic language scales, was significantly correlated with the Adaptive Behavior Assessment System–Second Edition Global Adaptive Composite, Conceptual, and Practical domains. Further investigation utilizing hierarchical regression analyses identified only the Speech scale on the CCC-2 as contributing unique variance to the prediction of adaptive behavior scores in the Global Adaptive Composite, Conceptual, and Practical domains.

Conclusion

Speech impairments may serve as flags to identify children with DS who are at risk for adaptive behavior deficits and reinforce the need for speech-language therapies that focus on speech for these individuals.

Supplemental Material

https://doi.org/10.23641/asha.13231985

Down syndrome (DS) is a neurodevelopmental disorder occurring in one of every 691 live births (Parker et al., 2010). DS is associated with global learning challenges as well as specific impairments in aspects of speech and language functioning that exceed general learning difficulties (Laws et al., 2000). Furthermore, as a disorder characterized by intellectual disability, DS is also associated with impairments in adaptive functioning (Tomaszewski et al., 2018). However, little is known about the linguistic correlates of impaired adaptive functioning in DS. Thus, the current investigation sought to examine how different speech and language skills (as measured by parent report) relate to adaptive behavior skills in young people with DS in order to help elucidate which speech and language abilities may be important early targets of treatment with the aim of improving later adaptive outcomes for this group.

Speech and Language Abilities in DS

Children with DS demonstrate relatively universal deficits in speech and language functioning relative to chronological age expectations, with greater impairments in structural (i.e., speech and syntax skills) than pragmatic (i.e., more social aspects of language, including nonverbal communication) language and expressive language compared to receptive language (McDuffie & Abbeduto, 2009). These deficits are apparent from early in development, as children with DS tend to produce first words much later than typically developing peers (Berglund, et al., 2001; Roberts et al., 2007) and learn new words at a slower rate (Berglund et al., 2001; McDuffie & Abbeduto, 2009).

Speech is an area of particularly significant impairment for youth with DS. Children with DS present with impaired abilities that are in excess of global learning challenges (Martin et al., 2009). These difficulties are thought to be due in part to the atypical craniofacial features of DS that impact oral motor development (Moura et al., 2008). However, these difficulties are not limited to the expressive domain. Youth with DS also present with impaired speech sound processing when evaluated receptively, as evidenced by performance on phonological awareness tasks that falls below mental age expectations (Bird et al., 2000; Cleland et al., 2009). Similarly, youth with DS have profound syntactic weaknesses compared to mental age–matched peers (Abbeduto et al., 2007).

Adaptive Behavior in DS

Adaptive abilities, the outcome of interest of this study, are broadly defined as an individual's functional skills surrounding developmentally appropriate, everyday activities. These skills can be further categorized into domains, including conceptual skills, social skills, and practical skills (American Psychiatric Association, 2013; Sparrow et al., 2008), which correspond to the three composites on the Adaptive Behavior Assessment System–Second Edition (ABAS-II; Burns, 2005), the instrument used in the current research. Individuals with DS tend to demonstrate relative strengths in socialization and more pronounced weaknesses in communication and practical or daily living skills (Fidler et al., 2006; Tomaszewski et al., 2018; Will et al., 2018). Throughout all stages of life, individuals with DS present with impairments in adaptive abilities (Dykens et al., 1994; Fidler et al., 2006; Will et al., 2018). Research assessing adaptive abilities in early development indicates that impairments across domains can be detected in the first 3 years of life (Fidler et al., 2006; Will et al., 2018). These challenges appear greater at older ages, as the development of adaptive skills in youth with DS does not keep pace with typically developing peers (Will et al., 2018).

Although the depth of impairment to speech-language and adaptive functioning in DS has been well researched, little is known about the linguistic correlates of adaptive behavior in youth with DS. This is surprising, as speech and language skills are required for the successful completion of many adaptive behaviors across different skill domains (e.g., calling for help when needed, asking for directions, inviting friends to one's home). In addition, in other neurodevelopmental disorders, different aspects of speech and language skills have been tied to adaptive function (Liss et al., 2001; Luyster et al., 2007). For example, in a sample of youth with autism spectrum disorder, vocabulary abilities were predictive of adaptive behaviors above and beyond the variance accounted for by IQ and autism symptomology (Liss et al., 2001). Thus, the current research sought to evaluate relationships between everyday structural and pragmatic language skills, as measured by the Children's Communication Checklist–Second Edition (CCC-2; Bishop, 2003) and adaptive behavior, measured using the ABAS-II, in a sample of individuals with DS.

Method

All research procedures were completed at the National Institutes of Health Clinical Center in Bethesda, MD. The study was approved by the National Institute of Health's Combined Neuroscience Institutional Review Board. See Lee et al. (2016) for details about the consent procedures.

Participants

Twenty-nine individuals with DS, including two siblings, participated in this study as part of a larger program of research at the National Institute of Mental Health (Lee et al., 2016). Participants were included in the current study if they had a medical diagnosis of DS and had complete ABAS-II, CCC-2, and nonverbal IQ data. The mean age of the sample was 13.46 years (range: 5–21), and 12 of the participants were male. The racial breakdown of the sample was as follows: Asian 10.3%, Black 6.9%, multirace 10.3%, and White 72.4%. Demographic information about the sample is summarized in Table 1.

Table 1.

Demographic characteristics of the sample and study assessment scores.

Demographic characteristics	M	SD	Range
Age (years)	13.46	5.12	5–21
Hollingshead SES	43.34	12.89	20–63
NVIQ standard score	56.83	16.23	24–95
ABAS-II
Global Adaptive Composite
Standard score	62.55	13.49	42–91
Raw score	351.93	86.44	189–552
Conceptual
Standard score	63.34	12.16	49–87
Raw score	108.79	34.15	53–178
Social
Standard score	79.76	13.91	58–104
Raw score	94.41	15.73	49–122
Practical
Standard score	58.41	16.53	40–90
Raw score	148.72	43.66	84–226
CCC-II
Structural Language Study Composite
Scaled score	4.47	2.19	1.25–9.75
Raw score	9.34	4.13	1.25–17.50
Pragmatic Language Study Composite
Scaled score	6.34	2.27	1.75–10.50
Raw score	7.47	3.40	1.25–16.00

Note. Standard scores: M = 100, SD = 15; scaled scores: M = 10, SD = 3. For the CCC-2, the study-derived Structural Language Composite presented is the mean of the following scales: Speech, Syntax, Semantics, and Coherence; the Pragmatic Language Scaled Score Composite is the mean of the following scales: Initiation, Scripted Language, Context, and Nonverbal Communication. For the ABAS-II, the Conceptual domain presented is the raw score sum of the following skill areas: Communication, Functional Academics, and Self-Direction (possible raw score range: 0–216); the Social domain presented is the raw score sum of the following skill areas: Leisure and Social (possible raw score range: 0–135); and the Practical domain presented is the raw score sum of the following skill areas: Community Use, Home Living, Health and Safety, and Self-Care (possible raw score range: 0–282). SES = socioeconomic status; NVIQ = nonverbal IQ; ABAS-II = Adaptive Behavior Assessment System–Second Edition; CCC-2 = Children's Communication Checklist–Second Edition.

Measures

The CCC-2 (Bishop, 2003) is a well-validated parent report questionnaire that assesses everyday speech and language skills in youth ages 4–16 years. It consists of 70 items within 10 scales that measure aspects of structural and pragmatic language. Structural language (i.e., the “form/content” of language; Bishop, 2015) is assessed through the Speech (e.g., articulation of sounds, speech fluency), Syntax (e.g., complexity of utterances, verb tense errors), Semantics (e.g., meaning and complexity of words and sentences), and Coherence (e.g., ability to clearly communicate) scales. The Initiation (e.g., social appropriateness of utterances), Scripted Language (e.g., repetitive language, impaired reciprocal use of language), Context (e.g., understanding different meanings of words/phrases, understanding humor), and Nonverbal Communication (e.g., appropriate eye contact, gestures) scales assess pragmatic language (Bishop, 2015). The CCC-2 has strong test–retest reliability (ranging from r = .86 to .96) and internal consistency (ranging from .94 to .96). It has been validated with clinical samples, including those with language impairments and autism spectrum disorders (McCauley, 2010; Volden & Phillips, 2010). In addition, the CCC-2 has been used as a measure of communication impairments in populations with DS in several prior studies (e.g., Laws & Bishop, 2004; Smith et al., 2017).

Raw scores from the CCC-2 were utilized in analyses. To reduce the number of analyses, structural language and pragmatic language composites were created, with the mean raw scores on the Speech, Syntax, Semantics, and Coherence scales comprising the structural language composite, and the mean raw scores on the Initiation, Scripted Language, Context, and Nonverbal Communication scales creating the pragmatic language composite (consistent with prior studies; Laws & Bishop, 2004; Philofsky et al., 2007). As higher scores denote poorer language skills, CCC-2 scores were reverse coded to allow for ease of interpretation. Thus, higher raw scores in the current investigation are indicative of stronger performance.

Lastly, it is important to note two participant characteristics relevant to the use of the CCC-2 with children and young adults with DS. First, the current study included nine young adults with DS (ages 17–21 years) who were outside of the normative age range of the CCC-2. As (a) we were utilizing raw scores for analyses, (b) no participants performed at ceiling level, and (c) the highest mental age of the young adults with DS in the sample was in the 8-year range, we deemed this break from normative procedures justified to help increase sample size and power. However, analyses were run with and without these participants to ensure rough comparability in findings (particularly pertaining to speech's association with adaptive behavior); see Supplemental Material S1. Second, nine participants in our sample had some hearing loss according to parent report. For these participants, the degree of hearing loss did not impact the ability to complete standard psychological testing procedures. Because the CCC-2 is not intended for use with individuals who have permanent hearing loss, we also ran the analyses without participants whose caregivers noted any hearing loss. As the pattern of findings, particularly speech's contribution to the ABAS-II, was similar (see Supplemental Material S2), we elected to report on analyses in the Results section that included all participants to maximize power and to describe findings for a representative sample of individuals with DS, given that about one third of individuals with DS have some hearing loss (Nightengale et al., 2017).

The ABAS-II (Harrison & Oakland, 2003) is a well-validated informant report questionnaire of everyday adaptive function for individuals ages 5–21 years (Burns, 2005). It is composed of an overall General Adaptive Composite (GAC) that can be divided into three theoretically derived domains that incorporate nine skill areas. The GAC has strong internal consistency (Cronbach's α = .98) as do each of the individual domains (Cronbach's α = .86–.93) and skill areas (Cronbach's α = .95–.97). The Conceptual domain encompasses the Communication (e.g., provides directions for multistep tasks, speech is clear), Functional Academics (e.g., follows common signs, composes letters/e-mails), and Self-Direction (e.g., continues working on difficult tasks, prioritizes academics over leisure) skill areas. The Social domain encompasses the Leisure (e.g., invites friends over, plans play activities) and Social (e.g., comments on others' emotions, maintains friendships) skill areas. Finally, the Practical domain encompasses the Community Use (e.g., asks store clerk for assistance, locates public restrooms), Home Living (e.g., prepares simple meals, keeps belongings tidy), Health and Safety (e.g., asks to see nurse when needed, follows rules related to safety), and Self-Care (e.g., fastens clothing, uses restroom independently) skill areas. Raw scores from the ABAS-II were utilized in analyses. It should be noted that the ABAS-II has a degree of measurement overlap with the CCC-2, as speech and language skills are embedded within certain items, particularly within the Conceptual domain's Communication skill area. To account for this overlap, the Conceptual domain of the ABAS-II will be assessed both at the domain level and with the Communication skill area removed (i.e., a composite of just the Functional Academics and Self-Direction skill areas).

The Differential Abilities Scale–Second Edition (Elliott, 2007 and Kaufman Brief Intelligence Test–Second Edition (Kaufman & Kaufman, 2004) were used to estimate nonverbal intellectual functioning for participants under the age of 17 years and older, respectively. Nonverbal IQ, ABAS-II, and CCC-2 scores are shown in Table 1.

Analytic Approach

To reduce the number of analyses completed, composite scores were created by averaging the raw scores for the Structural and Pragmatic Language subtests of the CCC-2 and creating raw score composites for ABAS-II domains by totaling skill area scores. To investigate associations between the raw CCC-2 composite scores and the raw ABAS-II domain scores, Pearson correlations were conducted. To streamline subsequent regression analyses, the scales of the structural or pragmatic language composites were only included as independent variables in follow-up hierarchical regression analyses if that composite was significantly correlated with at least one ABAS-II domain. For these analyses, Step 1 of the model included age, sex, and nonverbal IQ, whereas Step 2 included the individual scales of the CCC-2 from the relevant composite.

Results

Table 2 provides Pearson correlations for the CCC-2 and ABAS-II composites, age, sex, and nonverbal IQ. As seen in Table 2, only the structural language composite was significantly correlated with individual domains of the ABAS-II (i.e., the Conceptual and Practical domains) and the overall GAC. The pragmatic language composite was not significantly correlated with any domains of the ABAS-II or the GAC (ps > .05). Thus, regression analyses included only the CCC-2 scales that comprised the structural language composite (i.e., Speech, Syntax, Semantics, and Coherence). The correlations of the four CCC-2 structural language scales with the other variables described above are also included in Table 2.

Table 2.

Pearson correlations among variables of interest.

	1	2	3	4	5	6	7	8	9	10	11	12	13
1. Age	—	—	—	—	—	—	—	—	—	—	—	—	—
2. Sex a	.14	—	—	—	—	—	—	—	—	—	—	—	—
3. NVIQ	−.45*	−.20	—	—	—	—	—	—	—	—	—	—	—
4. ABAS GAC b	.66**	−.14	−.16	—	—	—	—	—	—	—	—	—	—
5. ABAS Con b	.63**	−.18	−.11	.97**	—	—	—	—	—	—	—	—	—
6. ABAS Soc b	.29	−.11	−.18	.69**	.60*	—	—	—	—	—	—	—	—
7. ABAS Prac b	.71**	−.10	−.17	.97**	.93**	.54**	—	—	—	—	—	—	—
8. CCC Struct Lang c	.33	−.25	−.09	.56**	.61**	.31	.51**	—	—	—	—	—	—
9. CCC Prag Lang c	.11	−.04	.41*	.29	.34	.22	.23	.58*	—	—	—	—	—
10. CCC Speech c	.38*	−.25	−.28	.62**	.65**	.38*	.59*	.91*	.34	—	—	—	—
11. CCC Syntax c	.17	−.41*	−.10	.40*	.47*	.14	.36	.89**	.36	.78**	—	—	—
12. CCC Semantics c	.28	−.08	.11	.36	.40*	.28	.31	.75**	.60**	.65**	.50**	—	—
13. CCC Coherence c	.27	.02	.07	.46*	.49**	.26	.42*	.76**	.81**	.55**	.58**	.50**	—

Note. See Table 1 for details about domains/scales that constitute the ABAS-II and CCC-2 composites. NVIQ = nonverbal IQ; ABAS = Adaptive Behavior Assessment System; GAC = General Adaptive Composite; Con = Conceptual; Soc = Social; Prac = Practical; CCC = Children's Communication Checklist; Struct Lang = Structural Language Study; Prag Lang = Pragmatic Language Study.

^a Sex: 0 = female; 1 = male.

^b Raw score.

^c Reverse-coded raw score.

^* p < .05.

^** p < .01.

To understand the contributions of the structural language scales to the concurrent prediction of the ABAS-II GAC and domains, hierarchical multiple regression analyses were completed using the individual scales of the structural language composite. The Speech scale of the CCC-2 predicted unique variance for the GAC, Conceptual, and Practical domains of the ABAS-II, but not the Social domain of the ABAS-II. No other scales of the CCC-2 (i.e., Syntax, Semantics, or Coherence) accounted for unique variance in either the GAC or individual ABAS-II domains. See Table 3.

Table 3.

Hierarchical regression model of ABAS-II GAC and domains for CCC-2 structural language scale scores.

	GAC a		Conceptual a		Social a		Practical a
Step 1 (Age, sex, NVIQ)	R 2Δ = .50; FΔ(3, 25) = 8.41***		R 2Δ = .49; FΔ(3, 25) = 8.02***		R 2Δ = .11; FΔ(3, 25) = 1.03.		R 2Δ = .56; FΔ(3, 25) = 10.72***
Step 2 (All variables)	R 2Δ = .19; FΔ(4, 21) = 3.10*		R 2Δ = .21; FΔ(4, 21) = 3.71*		R 2Δ = .15; FΔ(4, 21) = 1.09		R 2Δ = .17; FΔ(4, 21) = 3.22*
	β	t	β	t	β	t	β	t
Age	.60	3.96***	0.59	3.97***	.08	0.33	.70	4.96***
Sex b	−.10	−0.67	−.08	−0.55	−.25	−1.05	−.05	−0.35
NVIQ	.28	1.66	.34	2.01	−.14	−0.54	.35	2.19*
Speech c	.72	2.75*	.68	2.66*	.46	1.15	.72	2.97**
Syntax c	−.22	−0.98	−.09	−0.39	−.53	−1.54	−.18	−0.85
Semantics c	−.26	−1.43	−.27	−1.48	.10	0.34	−.35	−2.03
Coherence c	.13	0.76	.12	0.68	.24	0.91	.08	0.52

Note. ABAS-II = Adaptive Behavior Assessment System–Second Edition; GAC = General Adaptive Composite; CCC-2 = Children's Communication Checklist–Second Edition; NVIQ = nonverbal IQ.

^a Raw scores.

^b Sex: 0 = female; 1 = male.

^c Reverse raw scores.

^* p < .05.

^** p < .01.

^*** p < .001.

With regard to the connection between structural language and the Conceptual domain of the ABAS-II, one may wonder if this relationship was driven by the Communication subdomain (i.e., poor speech-language skills on one instrument predict poor speech-language skills on another instrument). To rule out this possibility, follow-up analyses investigated the impact of structural language scales on the Conceptual domain after the removal of the Communication skill area. Speech continued to account for unique variance in the Conceptual domain (β = .62, p = .03).

Discussion

The current paper examined whether parent-reported speech-language skills, as measured by the CCC-2, are significant concurrent predictors of adaptive behavior, as measured by the ABAS-II, in youth with DS. Study findings indicated that structural language skills, but not pragmatic language skills from the CCC-2, were significantly associated with global adaptive function abilities as well as the Conceptual and Practical domains of the ABAS-II. To further understand this relationship, the CCC-2 structural language scales were examined as predictors of adaptive function. For these analyses, the Speech scale was the only scale to predict unique variance in global adaptive function as well as in the Conceptual and Practical domains.

While the depth of impairment to speech-language and adaptative functioning in individuals with DS is well established, this study is the first to elucidate the association between these two constructs. This association is likely due to the central role that speech-language plays in the ability to functionally communicate in the community. As structural language difficulties persist into adulthood for individuals with DS (Rondal & Comblian, 1996), these findings have important implications for long-term adaptive outcomes in the population. Primarily, the results of this study suggest that speech skills may be important to consider when creating interventions that aim to improve adaptive functioning among individuals with DS.

Although individuals with DS often receive speech-language therapy as standard of care, this group may see adaptive skill benefit from speech-language interventions that target speech as well as other higher level forms of language. For example, interventions could work to support speech and language abilities in a manner that would allow one to ask family, friends, or acquaintances for help or receive assistance in the community. These interventions may also support social skills as individuals learn how to initiate conversations and invite peers to engage in leisure activities. Lastly, improving speech may assist with the completion of adaptive behaviors relevant to the work environment, such as communicating with customers at a store or employers. Speech-language interventions occurring early in life, such as milieu communication teaching, palatal plate therapy, and broad target speech recasts, show promise in improving speech in DS populations (Yoder et al., 2016). Thus, future research may benefit from examining the cascading effects of these interventions on adaptive functioning skills in populations with DS. Furthermore, as speech-language difficulties are likely to improve but persist throughout the life span, augmenting adaptive skills should be a priority for intervention, as deficits in these skills have implications for various functional outcomes, including occupational status and residential independence, in individuals with DS (Foley et al., 2013; Woolf et al., 2010). As the life span for individuals with DS has grown considerably, adults with DS have the potential to outlive caregivers. Thus, identifying ways to improve adaptive skills in this population should be at the forefront of intervention research. Finally, these findings may transfer to other syndromic forms of intellectual disability that experience difficulties in speech or language skills (e.g., fragile X syndrome, Prader–Willi syndrome; Reus et al., 2011; Zingerevich et al., 2009).

We now turn to the study's limitations and directions for future research. Like many studies of individuals with DS, our sample size was small and we were underpowered to detect small to medium effect sizes. Thus, future research may benefit from combining samples across sites to examine whether these relations hold true for larger samples. In addition, as this study was cross-sectional in design, future research should look to employ longitudinal approaches to further elucidate the relationship between structural language and adaptive skills over time. Lastly, since this study examined only communication abilities as a predictor of adaptive function in youth with DS, we were not able to speak to other cognitive abilities (e.g., executive function) that contribute to adaptive behavior in this group. Thus, future research involving larger samples should investigate how different cognitive domains relate to adaptive behavior. Such research may further our understanding of the cognitive underpinnings of everyday functioning in individuals with DS and inform treatment approaches that aim to improve functional outcomes for this group.

Author Contributions

Liv Clasen: Writing – review & editing (Equal). Elizabeth Adeyem: Writing – review & editing (Supporting). Nancy Raitano Lee: Formal analysis (Equal), Investigation (Supporting), Supervision: (Lead), Writing – original draft (Supporting), Writing – review & editing (Supporting).

Funding Statement

Funding for this study was provided to Jay N. Giedd (protocol PI) at the time data were collected by the National Institute of Mental Health (1 ZIA MH002794-13; Protocol ID 89-M-0006).