Correlates of adaptive behavior profiles in a large cohort of children with autism: The autism speaks Autism Treatment Network registry data

Abstract

Children with autism spectrum disorder have deficits in adaptive functioning. This study examines the adaptive behavior, its association with cognitive ability, gender, age, and symptom severity in children with autism spectrum disorder. Using data from Autism Treatment Network registry, the adaptive behavior profiles were examined in 2538 school-aged children (between 5 and 17 years, mean: 8.8 years, standard deviation: 3.0) who had an overall intelligence quotient and Vineland Adaptive Behavior Scale scores available. The children were grouped according to their intelligence quotient (low intelligence quotient < 70; borderline intelligence quotient = 70–85; average intelligence quotient > 85), age (5–10 and 11–17 years), and gender for the analyses. Significantly lower Vineland Adaptive Behavior Scale scores were found in borderline and average intelligence quotient groups when compared to mean intelligence quotient, while an opposite pattern was seen in the low intelligence quotient group, with better adaptive behavior scores than mean intelligence quotient. Vineland Adaptive Behavior Scale standard scores were positively correlated with intelligence quotient and poorly associated with autism spectrum disorder severity. Younger children had significantly higher Vineland Adaptive Behavior Scale scores. Adjusted comparisons by gender were not significant. Adaptive behavior profiles in the intelligence quotient categories are discussed. This study confirms a positive relationship between adaptive behavior and intellectual function in autism and indicates that children with higher intelligence quotient and older age are specifically impaired, with lower adaptive behavior, highlighting the need for assessment and targeted intervention in these groups. Future directions for research are discussed.

Keywords

adaptive behavior age autism spectrum disorder cognitive ability gender intellectual functioning sex differences

Introduction

Children with autism spectrum disorder (ASD) often have deficits in their cognitive functioning and adaptive behavior (Carter et al., 1998; Klin et al., 2007), and there is wide variability in intellectual functioning in ASD population. Contrary to the previous notion that almost 75% of individuals with ASD have intellectual disability, recent epidemiological studies have indicated that about half of the individuals with ASD present with intelligence quotient (IQ) of 70 or more (Chakrabarti and Fombonne, 2005; Charman et al., 2011). Children with autism without comorbid intellectual disability demonstrate relative strengths in cognitive and formal language abilities with an expectation to achieve positive outcomes. Intellectual functioning has been found to be the best prognostic indicator of outcomes in ASD (Eaves and Ho, 2008). However, within the normal IQ range, the outcomes remain variable (Howlin et al., 2004).

Adaptive behavior is defined as the collection of conceptual, social, and practical skills that have been learned and are performed in everyday lives (American Psychiatric Association (APA), 2013; Schalock et al., 2010). The Vineland Adaptive Behavior Scale (VABS) is the most commonly used scale to measure the adaptive behaviors in individuals with ASD (Gillham et al., 2000; Sparrow et al., 2005). Children with ASD display poorer adaptive behavior when compared to typically developing children and children with other developmental disorders (Carter et al., 1998; Mouga et al., 2015). Adaptive behavior has been observed to be more strongly related to optimal outcomes in adulthood than cognitive function (Farley et al., 2009). Hence, adaptive behavior is an important component in comprehensive evaluation of ASD and a determinant of prognosis (Klin et al., 2007; Sparrow et al., 2005; Tomanik et al., 2007). Consensus is building regarding the importance of early intervention in ASD to improve long-term outcomes, and adaptive behavior presents as an achievable and crucial intervention target (Bal et al., 2015; Howlin, 2013).

A heterogeneous adaptive behavior profile in autism has been explored in many studies although evidence regarding a distinct autism-specific adaptive functioning profile is mixed. Furthermore, this profile may be impacted by intellectual functioning (Bölte and Poustka, 2002; Liss et al., 2001). A typical “autism profile” on VABS has been reported with higher scores for motor and daily living skills, lowest scores for socialization, and intermediate scores for communication (Carter et al., 1998; Kraijer, 2000). In a large sample of cognitively able children with ASD (IQ > 70), greatest impairments were seen in socialization skills, with moderate delays in communication and daily living skills (Kanne et al., 2011). In younger children, Perry et al. found that such profile emerged when age equivalents were used at different cognitive levels but not when standard scores were examined (Perry et al., 2009). A typical autism profile of adaptive behavior was not observed in younger children with intellectual disability (Fenton et al., 2003). In another recent study, this profile was not supported, and a relative weakness in daily living skills across different intellectual ability groups was found (Matthews et al., 2015). Further evidence to support or refute the existence of typical ASD adaptive behavior profiles in a large ASD sample with varied cognitive abilities and age range will be helpful in determining intervention needs, guiding outcome measures, and establishing the role of adaptive behavior in diagnostic dilemmas for subtle cases (Tomanik et al., 2007).

Previous research has identified several variables associated with adaptive behavior in ASD. A positive correlation exists between adaptive behavior and intellectual ability (Bölte and Poustka, 2002). While adaptive behavior profiles are commensurate with cognitive ability in typically developing children, individuals with ASD demonstrate a wide gap between IQ and adaptive behavior, particularly in intellectually able individuals, highlighting specific deficits associated with ASD (Bölte and Poustka, 2002; Liss et al., 2001; Lopata et al., 2013). Age has also been investigated in relation to adaptive behavior in ASD. In cross-sectional studies, a strong negative correlation between age and adaptive behavior has been reported (Kanne et al., 2011; Klin et al., 2007; Pugliese et al., 2015). Longitudinal analyses indicate that the adaptive skills acquisition rate is attenuated among individuals with ASD when compared to typically developing peers (Gabriels et al., 2007). Some researchers have not found an association between age and adaptive behavior in individuals with high-functioning ASD (Kenworthy et al., 2010). Similar varying trends exist in association between adaptive behavior and ASD symptom severity. In older children, a weak relationship between symptoms and adaptive behavior has been reported (Kanne et al., 2011; Klin et al., 2007), while in younger children, a moderate to strong association between ASD severity and adaptive behavior has been reported (Perry et al., 2009). It is important to investigate the effects of age, IQ, and ASD severity on adaptive behavior in a large sample of children with varying age range and cognitive ability in order to establish the association between these variables and adaptive behavior and to reconcile inconsistent findings. Understanding the variables relating to adaptive behavior in ASD is critical, given the potential association between adaptive skills and long-term ASD outcomes (Farley et al., 2009).

ASD has a high male-to-female ratio, with an average of 3:1 and rising to 10:1 in high-functioning individuals as compared to 2:1 in ASD with coexistent ID (Fombonne, 2009; Loomes et al., 2017). This identifies IQ as a potential confounding variable when examining sex differences in ASD, hence the need to control its effects (Volkmar et al., 1993). A study comparing boys and girls with high levels of autistic traits, who did or did not meet diagnostic criteria for ASD, suggested that girls are more likely to get a diagnosis when they exhibit behavioral problems or low cognitive ability (Dworzynski et al., 2012). Concomitantly, girls are reported to “camouflage” ASD symptoms, and high-functioning girls with ASD are often underdiagnosed (Dean et al., 2016; Shattuck et al., 2009). Very few studies have looked into sex differences in adaptive behavior in ASD. A study on a large sample of children with ASD suggested weaker adaptive skills and lower cognitive ability in females. This sample consisted of a lower proportion of females with average intelligence. IQ reductions mediated reduced adaptive behavior in females when compared to males (Frazier et al., 2014). It is unknown if sex differences in adaptive behavior profiles persist when controlling for IQ. Such information may help build evidence to standardize symptom ascertainment based on gender, a practice that is recommended but not often utilized in current clinical practice (Constantino and Charman, 2016).

This study aims to build upon existing knowledge regarding adaptive behavior in children with ASD by examining data from a large well-defined cohort of children—the Autism Speaks Autism Treatment Network (ATN) registry. The study aims to investigate (1) the association between adaptive behavior function and overall IQ in children with varied cognitive functioning. Based on previous studies, we hypothesize that there is a positive relation between adaptive behavior and cognitive function; (2) the discrepancies between adaptive and cognitive function across the range of cognitive abilities. We hypothesize that the discrepancy is more apparent in higher functioning children and adolescents; (3) the differences in adaptive behavior in relation to age and severity. Based on previous studies, we hypothesize that older children have more deficits in adaptive behavior, and ASD severity is poorly associated with adaptive behavior function; (4) sex differences in the adaptive behavior of IQ-matched children. We hypothesize that girls have poorer adaptive skills than boys; (5) if a typical “autism profile” for adaptive behavior exists in our large study cohort. Characterizing the adaptive behavior profiles in ASD is essential to the development of customized intervention strategies to optimize self-sufficiency for autistic individuals at all levels of functioning.

Methods

Participants

The ATN registry includes children and adolescents with ASD enrolled across 14 sites in the United States and Canada. The study sample included 2538 children, enrolled between 2008 and 2016, ages 5–17 years (mean: 8.8, standard deviation (SD) = 3.0), who had IQ and any VABS subscale or total standard scores available. For age-specific analysis, they were divided into two groups: young (5–10 years, n = 1939) and old (11–17 years, n = 599). Other variables considered for analysis were race (White vs non-White), ethnicity (Hispanic vs non-Hispanic), and primary caregiver education level (at least some college vs at most high school). One of the ATN registry inclusion requirements was parent or guardian’s fluency in English or Spanish. Exclusion criteria included medical conditions precluding valid testing. The study sample consisted of 85% boys (boys vs girls’ ratio 5.6:1), Caucasian (83%), Asian (4%), Black or African (7%), multiracial (6%), and a majority of non-Hispanic (91%) children. Autism was diagnosed in 62% (n = 1533), Asperger’s in 15% (n = 374), and pervasive developmental disorder not otherwise specified (PDD-NOS) in 23% (n = 555). Most of the primary caregivers (82%) received higher education (at least some college or more).

Measures

Adaptive functioning

The VABS second edition (VABS-II) is a standardized, structured parent/caregiver survey interview of adaptive behaviors and has been shown to have strong reliability and validity (Sparrow et al., 2005). The ATN uses the Vineland-II, Survey Interview Form. This measure produces an overall composite standard score and age-equivalent standard scores in four domains: communication, daily living skills, socialization, and motor skills. In this study, standard scores for communication, daily living skills, and socialization domains along with the composite standard scores were used. The motor skills domain was not included in the analyses, as it is only administered to individuals under the age of 6 years.

Cognitive ability

A number of standardized tests with established psychometric properties were used to obtain full scale or abbreviated IQ scores, a common scenario in autism practice and research. These included Stanford–Binet (Roid, 2003), Mullen Scales of Early Learning (Mullen, 1995), Wechsler Abbreviated Scale of Intelligence (WASI) (Wechsler, 1999), Differential Abilities Scale-second edition (DAS-II) (Elliott et al., 1990), Wechsler Intelligence Scale for Children-fourth edition (WISC-IV) (Wechsler, 2003), or Wechsler Preschool and Primary Scale of Intelligence-third edition (WPPSI-III) (Wechsler, 2002), depending on the ATN site’s preference, scale availability, and child’s age and language ability. Most participants were administered the Stanford–Binet (n = 2024) followed by WISC-IV (n = 158), DAS-II (n = 157), Mullen (n = 99), WPPSI (n = 56), and the WASI (n = 44). For all participants, norm-referenced overall deviation scores were used as a measure of IQ. Children were divided into three IQ categories based on SD below the population mean: low (IQ < 70, n = 838), borderline (IQ between 70 and 85, n = 506), and average (IQ > 85, n = 1194) intellectual functioning.

Autism diagnosis

Assessment of ASD was based on the Diagnostic and Statistical Manual of Mental disorders (4th ed., text rev.; DSM-IV-TR) criteria using the checklist that allowed the clinician to indicate the diagnosis of autism, Asperger’s, or PDD-NOS. The majority (88%) of the children were assessed using Autism Diagnostic Observation Schedule (ADOS) or its second edition (ADOS-2) (Lord et al., 2012). Severity of symptoms was assessed using ADOS/ADOS-2 calibrated severity scores.

Participants were eligible to be included in the database after meeting ASD diagnostic criteria, as determined by clinical consensus using ADOS, ADOS-2, and/or DSM-IV-TR (APA, 2000).

Statistical methods

IQ, demographics, DSM-IV-TR diagnosis, calibrated ADOS severity score, and VABS total and subscale standard scores were tabulated for categorical variables and described with means and SDs for continuous variables. Continuous age and IQ variables were converted into categories based on clinically meaningful groups, and the use of categorical variables facilitated interpretation of interactive effects. Bivariate tests of baseline characteristics with each VABS score were performed using t-tests or analysis of variance (ANOVA) for categorical variables and Spearman correlation for continuous variables. Cohen’s d and omega-square effect sizes were reported for t-tests and ANOVA, respectively. Covariates for multiple linear regression models were selected based on a 0.05 significance threshold from the bivariate tests. Tukey’s post hoc pairwise comparison tests were used to assess differences in scores by domain category within each IQ category.

The associations between VABS standard scores and IQ (continuous) were assessed using Pearson correlation and scatterplots, and differences in the VABS scores and IQ (VABS score–IQ) were described in each IQ category using means, SDs, and boxplots and were tested using paired t-tests. To determine the association between adaptive behavior and cognitive intelligence, each regression model used VABS score as the dependent variable and IQ category, with reference group >85, as the main independent variable of interest, while controlling for covariates. Adjusted least square means, differences in the adjusted means (unstandardized and standardized beta coefficients), standard errors, and p values are reported. Listwise deletion was used in the multivariable models, and multiple imputation sensitivity analyses were also performed. A Bonferroni-corrected significance threshold of 0.0125 (=0.05/4) was used to account for the four model outcomes. To determine if the effect of IQ on adaptive behavior differs by gender, we assessed the IQ category × gender interaction in each multivariable model. Likewise, to determine if the effect of IQ on adaptive behavior is modified by age (5–10 vs 11–17 years), we assessed the IQ category × age group interaction in each model. The associations between adaptive behavior and age or gender were assessed using adjusted regression models overall and within each IQ group, and the association between adaptive behavior and IQ was assessed within each age group.

Results

Less than 10% of data were missing for each covariate, except for ADOS severity score, which was missing for 12%. Table 1 indicates that the VABS scores for adaptive behavior composite and the four subscales differ significantly by IQ category, age group, DSM-IV diagnosis, race, primary caregiver education, and calibrated ADOS severity score (negative correlation) on bivariate analyses. Of these variables, IQ, age, and DSM-IV diagnosis were significantly correlated with VABS scores (except for small effect size for socialization scores and DSM-IV diagnosis). The pattern of VABS profile differed in the IQ categories. In the borderline and average IQ groups, both daily living skills and communication scores differed significantly from socialization score, while in the low IQ group none of the pairwise comparisons were significant. There was no significant difference between the daily living skills and communication scores in any of the IQ categories.

Table 1.

VABS standard scores overall and by IQ and baseline covariates.

Variable^a	Overall % (n) or mean ± SD	Adaptive behavior composite	Communication	DLS	Socialization
Overall sample (n = 2538)	Overall % (n) or mean ± SD	72.1 ± 12.1 (n = 2455)	75.7 ± 14.6 (n = 2535)	76.0 ± 14.8 (n = 2534)	69.7 ± 12.6 (n = 2532)
IQ		***	***	***	***
<70	33.0% (838)	63.1 ± 10.6	64.3 ± 11.9	65.9 ± 13.0	62.5 ± 11.3
70–85	19.9% (506)	72.9 ± 8.9	76.4 ± 9.8	77.2 ± 11.9	71.6 ± 11.3
>85	47.0% (1194)	78.0 ± 10.4	83.5 ± 12.8	82.6 ± 13.1	73.8 ± 11.7
Omega-square		0.29	0.33	0.25	0.16
Age		***	***	***	***
5–10 years	76.4% (1939)	74.0 ± 11.6	77.9 ± 14.5	77.5 ± 14.3	71.7 ± 11.9
11–17 years	23.6% (599)	65.9 ± 11.7	68.8 ± 12.9	70.9 ± 15.1	62.9 ± 12.3
Cohen’s d		0.69	0.64	0.46	0.74
Gender		NS	NS	NS	NS
Female	15.2% (386)	71.7 ± 12.3	75.4 ± 14.9	75.1 ± 14.7	69.7 ± 12.5
Male	84.8% (2150)	72.1 ± 12.1	75.8 ± 14.6	76.1 ± 14.8	69.6 ± 12.6
Cohen’s d		0.03	0.02	0.07	0.00
DSM-IV diagnosis		***	***	***	***
Autism	62.3% (1533)	68.8 ± 11.7	72.0 ± 14.1	72.4 ± 14.2	66.8 ± 12.3
PDD-NOS	22.5% (555)	75.8 ± 10.9	79.4 ± 12.8	79.7 ± 13.2	74.5 ± 11.2
Asperger’s	15.2% (374)	79.7 ± 10.9	85.3 ± 13.4	85.5 ± 13.6	74.3 ± 12.0
Omega-square		0.12	0.11	0.09	0.01
Race		***	***	***	***
White	82.8% (2008)	72.6 ± 12.1	76.3 ± 14.6	76.5 ± 14.9	70.1 ± 12.5
Non-White	17.2% (416)	69.8 ± 12.3	73.2 ± 14.8	73.9 ± 14.2	67.7 ± 12.6
Cohen’s d		0.23	0.21	0.18	0.19
Primary caregiver education		***	***	***	***
At most HS	18.3% (433)	68.6 ± 11.7	70.6 ± 13.3	73.1 ± 14.3	67.0 ± 12.1
Some college or more	81.7% (1937)	73.0 ± 12.1	77.0 ± 14.7	76.8 ± 14.8	70.3 ± 12.6
Cohen’s d		0.36	0.44	0.25	0.26
ADOS severity score^b (n = 2222)	7.2 ± 2.0	r = −0.18***	r = −0.15***	r = −0.18***	r = −0.18***

VABS: Vineland Adaptive Behavior Scale; IQ: intelligence quotient; DLS: daily living skills; SD: standard deviation; ANOVA: analysis of variance; DSM: Diagnostic and Statistical Manual of Mental Disorders; PDD-NOS: pervasive developmental disorder not otherwise specified; HS: high school; ADOS: Autism Diagnostic Observation Schedule; NS: not significant.

VABS standard scores reported with mean ± SD for each category and tested using ANOVA.

Spearman correlation coefficient r reported for calibrated ADOS severity score.

Cohen’s d and omega-square: bold values indicate more than medium effect size (Cohen’s d ⩾ 0.5; omega-square ⩾ 0.06).

***

p < 0.001.

Table 2 summarizes the results from the multivariable regression analyses of the VABS scores, adjusting for the covariates mentioned above. IQ and age of the children were strongly correlated with the VABS scores. Comparing multiple imputation to listwise deletion, none of the results differed in terms of significance, and all estimates from both methods of handling missing data were similar. Results are reported from the models that used listwise deletion.

Table 2.

VABS standard score results from multivariable regression models.

Dependent variable	Independent variable	Category	Adj.^a LS mean	Difference	Std. beta	SE	p value
Adaptive behavior composite	IQ	<70	62.04	−12.62	−0.49	0.56	<0.0001*
		70–85	70.51	−4.15	−0.14	0.61	<0.0001*
		>85 (Ref)	74.66
	Age	5–10 years	73.41	8.67	0.28	0.56	<0.0001*
	Age	11–17 years (Ref)	64.74
	Gender	Female	69.54	0. 58	0.02	0.62	0.3509
	Gender	Male (Ref)	68.96
Communication	IQ	<70	62.17	−17.03	−0.55	0.63	<0.0001*
		70–85	72.97	−6.22	−0.17	0.69	<0.0001*
		>85 (Ref)	79.20
	Age	5–10 years	76.48	10.07	0.28	0.64	<0.0001*
	Age	11–17 years (Ref)	66.41
	Gender	Female	71.84	0.49	0.01	0.70	0.4877
	Gender	Male (Ref)	71.36
DLS	IQ	<70	66.56	−13.91	−0.45	0.70	<0.0001*
		70–85	76.66	−3.81	−0.10	0.77	<0.0001*
		>85 (Ref)	80.47
	Age	5–10 years	77.85	6.57	0.18	0.71	<0.0001*
	Age	11–17 years (Ref)	71.28
	Gender	Female	74.31	−0.31	−0.01	0.78	0.6867
	Gender	Male (Ref)	74.62
Socialization	IQ	<70	61.02	−8.99	−0.33	0.63	<0.0001*
		70–85	68.59	−1.43	−0.04	0.69	0.0375
		>85 (Ref)	70.01
	Age	5–10 years	71.26	9.43	0.30	0.63	<0.0001*
		11–17 years (Ref)	61.82
	Gender	Female	67.15	0.76	0.02	0.69	0.2739
		Male (Ref)	66.40

VABS: Vineland Adaptive Behavior Scale; IQ: intelligence quotient; Adj.: adjusted; LS mean: least squares mean; Std.: standardized; SE: standard error; DLS: daily living skills; Ref: reference group.

Multivariable model includes IQ category, age group, DSM-IV diagnosis, race, ethnicity, primary caregiver education, and ADOS severity score; gender was not included as a covariate for the adjusted IQ and age estimates.

Significant at the Bonferroni-corrected 0.0125 level.

Figure 1 shows the positive correlation of IQ with VABS composite (r = 0.58, p < 0.0001), communication (r = 0.64, p < 0.0001), daily living skills (r = 0.52, p < 0.0001), and socialization (r = 0.42, p < 0.0001). Children with low and borderline IQ had lower scores for VABS composite (low: mean, 62.04, p < 0.0001; borderline: mean, 70.51, p < 0.0001; average: mean, 74.66) and four subscales (p < 0.0001, ω² = 0.22) when compared with children with average IQ. Figure 2 illustrates the gap between IQ and adaptive behavior in each IQ category. The VABS composite, communication, socialization, and daily living skills scores exceeded mean IQ in low IQ group (p < 0.0001 for each difference, Cohen’s d range: 0.86–1.07), whereas the VABS composite, communication, and socialization scores were below mean IQ in both the borderline and average IQ groups (p < 0.0001 for each difference, Cohen’s d range borderline: 0.47–0.49; average: 1.35–1.85; except communication in borderline group, p = 0.0155,Cohen’s d: 0.11); and daily living skills score was lower than IQ in the IQ > 85 category (p < 0.0001, Cohen’s d: 1.27) (Table 3). Although adaptive behavior improved with increasing IQ, the discrepancy between the adaptive behavior and IQ widened in children with average intellectual functioning (IQ > 85).

Figure 1.

Scatterplots of VABS standard scores and IQ.

Figure 2.

VABS standard scores and IQ by IQ category.

Table 3.

Differences in VABS standard scores and IQ by IQ category.

Variable^a	IQ	Adaptive behavior composite-IQ	Communication-IQ	DLS-IQ	Socialization-IQ
IQ (n = 2538)	81.3 ± 24.1
<70 (n = 838)	52.8 ± 8.0	10.0 ± 10.3***	11.4 ± 10.7***	13.1 ± 12.3***	9.7 ± 11.2***
70–85 (n = 506)	77.5 ± 4.7	−4.6 ± 9.9***	−1.1 ± 10.5*	−0.35 ± 12.8	−6.0 ± 12.2***
>85 (n = 1194)	102.9 ± 11.8	−25.0 ± 13.7***	−19.5 ± 14.4***	−20.4 ± 16.0***	−29.1 ± 15.7***

VABS: Vineland Adaptive Behavior Scale; IQ: intelligence quotient; DLS: daily living skills; SD: standard deviation.

VABS standard scores-IQ differences reported with mean ± SD for each category and tested using paired t-test.

***

p < 0.001, *p < 0.05.

Age and adaptive behavior were negatively correlated on all subscales. Adjusted comparisons of VABS scores by age group were significant (p < 0.0001 for each subscale score; ω² range: 0.04–0.12), with younger children having better composite, communication, daily living skills, and socialization skills. In the adjusted linear regression model, controlling for IQ and other variables, the mean VABS composite score among younger children was 8.7 units higher than among older children (p < 0.0001, ω² = 0.11). Similar results were found on subgroup adjusted analysis for younger versus older children within each IQ category (p < 0.0001 for each difference; low IQ: ω² range, 0.05–0.10; borderline IQ: ω² range, 0.02–0.10; average IQ: ω² range, 0.05–0.17), except for daily living skills in borderline group (p = 0.0032, ω² = 0.02). In older children, VABS scores were significantly different when comparing the low and average IQ groups (p < 0.0001; ω² = 0.07–0.21) but not when comparing borderline and average IQ groups (p > 0.0125 for each score difference; ω² = 0.00–0.01), indicating that IQ was not significantly associated with VABS scores among older children with IQ > 70. Figure 3(a) shows that the effect of IQ on adaptive behavior composite and communication scores differed by age; however, the effect size was small. Age did not modify the effect of IQ on daily living or socialization scores.

Figure 3.

(a) Age interactions with IQ on VABS composite and subscales standard scores and (b) gender interactions with IQ on VABS composite and subscales standard scores.

There was no association between gender and VABS scores on the bivariate or multivariable analyses. However, in the subgroup adjusted analyses within each IQ category with significance threshold 0.05, females with IQ > 85 had higher adaptive scores than IQ-matched males for composite (adjusted mean: female 76.4 vs male 74.2, p = 0.0233, 95% confidence interval (CI) for difference (0.31, 4.07)) and communication (adjusted mean: female 81.3 vs male 78.6, p = 0.0199, 95% CI for difference (0.42, 4.84)). When applying the Bonferroni-corrected significance threshold of 0.0125, none of these mean differences were significant. Figure 3(b) shows that the effect of IQ on VABS composite, communication and daily living scores differed by gender; however, the effect size was small. Gender did not modify the effect of IQ on socialization score.

While ADOS severity scores and VABS scores were negatively correlated (Table 1), the multivariable models did not show a significant association between ADOS severity and VABS scores. In Table 2 models, higher primary caregiver education was associated with higher composite (β = 2.03, SE = 0.57, p = 0.0004) and communication (β = 3.30, SE = 0.65, p < 0.0001) scores but not with socialization (β = 1.46, SE = 0.64, p = 0.0226) or daily living skills scores (β = 1.16, SE = 0.72, p = 0.1069). On multivariable analysis, race (White vs non-White) was not independently associated with VABS scores (β = 0.01, SE = 0.59, p > 0.0125 for composite), while non-Hispanic children had better scores in the communication (β = 0.05, SE = 0.93, p < 0.0125) and socialization (β = 0.06, SE = 0.92, p < 0.0125) domains but not for composite (β = 0.04, SE = 0.82, p = 0.0194), and daily living (β = 0.03, SE = 1.04, p = 0.0837).

Discussion

The study describes adaptive behavior profiles using the ATN registry, one of the largest ASD datasets that include school-aged children with varying cognitive abilities. We found a strong positive correlation between adaptive behavior and cognitive function. Children with higher IQ had more discrepancies between adaptive behavior and cognitive function compared to the children with intellectual disability. Age was negatively correlated with adaptive behavior and ADOS severity was poorly associated with adaptive behavior. Additionally, there was no association between gender and overall adaptive behavior. A typical autism profile was not observed as significantly lower socialization scores were seen only in groups with IQ > 70, with no significant difference between daily living and communication scores.

Our study found a positive correlation between cognitive and adaptive function across different cognitive level groups. This confirms previous findings that overall IQ is strongly associated with adaptive behavior function in children with ASD, even when controlling for ADOS severity (Kanne et al., 2011; Liss et al., 2001; Perry et al., 2009). We found that IQ had the strongest correlation with communication (r = 0.64) and the weakest with socialization (r = 0.42) scores (Kanne et al., 2011). Strong correlation between the communication scores and IQ is not surprising as the IQ tests have verbal subscales, and good language skill is helpful in engagement during the assessment. Additionally, the study also found that the children with higher IQ had more discrepancies between adaptive behavior and cognitive ability compared to the children with intellectual disability, for whom adaptive skills are relative strengths in relation to their cognitive functioning (Bölte and Poustka, 2002; Liss et al., 2001; Perry et al., 2009). To quantify the deficit, we found that the mean adaptive composite and subscale scores in those with average intelligence (IQ > 85) fall between one to two SD below the population mean IQ. This implies that the children without comorbid intellectual disability experience substantial challenges in everyday life and are not consistently able to apply their intellectual abilities into functional skills.

We observed a negative correlation between age and adaptive behavior and a poor association between adaptive behavior and ASD symptoms (Kanne et al., 2011; Klin et al., 2007; Pugliese et al., 2015). We used the age groups 5–10 years and 11–17 years to corroborate our results with the developmental level and functional requirement (primary school vs middle-high school age). The negative correlation between age and adaptive behavior prevailed in children with or without intellectual disability. We observed significant differences in adaptive behavior and cognitive function in older children with IQ > 70 that is, children who were cognitively higher functioning. This attenuation in adaptive behavior was most marked for socialization skills (Figure 3(a)). As children reach the adolescent years, environmental expectations increase, and the level of intellectual ability required to successfully learn adaptive skills in order to cope with environmental demands is much higher; and this may not be attained in cognitively able (having IQ > 70) adolescents with ASD. Moreover, children with ASD without comorbid intellectual disability are often diagnosed later than the cognitively impaired children (Shattuck et al., 2009). Hence, our findings identify intellectually able adolescents at relatively high risk for adaptive behavior impairments and emphasize the need for careful assessment in children with late ASD diagnosis and to focus on outcome research in this subgroup.

Our fourth hypothesis that the girls with ASD have poorer adaptive skills than boys was only accepted in the low IQ group; however, the effect size was extremely small. While girls with borderline IQ had similar adaptive scores to the IQ-matched males, females with average IQ were observed to have slightly better mean adaptive scores than their male counterparts, but a statistical significance was not achieved when the Bonferroni-corrected significance threshold of 0.0125 was applied. Additionally, we did not observe a significant difference in adaptive behavior standard scores with IQ reduction in girls when compared to boys. To our knowledge, this is the first study describing the sex differences in adaptive behavior profiles in children with ASD in relation to their cognitive abilities. A poorer cognitive and adaptive skills in females have been shown previously (Frazier et al., 2014). Our results are limited for interpretation, as we did not find marked differences in adaptive function between IQ-matched males and females with an ASD diagnosis. Our cohort had higher proportion of males (boys vs girls’ ratio 5.6:1), which may have implication on generalizability of the results. However, it is discernible that impairment in adaptive function in ASD is consistent, irrespective of gender, and it is an important outcome measure for consideration in all children with ASD.

Finally, we found the socialization scores to be significantly lower than the communication and daily living scores, yet the communication and daily living skills did not differ significantly in borderline and average IQ groups (IQ > 70). In the low IQ group, no significant differences were seen between the domains. Hence, existence of a typical autism profile for adaptive behavior was not established in our large cohort although we found relatively more significant impairment in socialization skills in children with IQ > 70. Our results are consistent with other studies that have shown significant impairment in socialization skills while describing adaptive behavior profiles in individuals with ASD and high IQ (Kanne et al., 2011; Klin et al., 2007; Liss et al., 2001; Lopata et al., 2013). For communication and daily living skills, we did not find any significant difference in any of the IQ categories in contrast to some previous studies (Carter et al., 1998; Kraijer, 2000). Nonetheless, children with ASD without comorbid intellectual disability exhibit significant delays in all the domains of adaptive behavior relative to their cognitive function. In our cohort, a statistically significant difference was not found between any adaptive behavior domain in children with low IQ (<70). This is in contrast to the findings of Matthews et al. (2015), who observed daily living skills to be the lowest and socialization skills to be higher than communication in children with ASD and coexistent intellectual disability (Matthews et al., 2015). Previous smaller studies that have compared the VABS profiles in children with intellectual disability with or without autism have shown that the adaptive behavior does not follow the typical autism profile in children with autism and intellectual disability (Fenton et al., 2003; Liss et al., 2001). Our study confirms and extends this important finding to a broader age group and to children with ASD without coexistent intellectual disability.

To our knowledge, this is the largest sample, comprising children and adolescents with ASD analyzed for the adaptive behavior profiles. Our findings not only add to the growing literature on adaptive behavior in ASD, but also have implications on current clinical practice. It is evident that adaptive behaviors are specifically impaired in children with ASD without intellectual disability and should be included when characterizing impairment in ASD. These children are often considered to have a milder level of disability (“high functioning”), with an expectation of positive outcomes based on their higher cognitive abilities. It is particularly concerning that these children may not receive targeted interventions focusing on adaptive behaviors due to their cognitive status. This study highlights the importance of the assessment of adaptive behavior in children with ASD, irrespective of cognitive ability, in order to provide interventions with explicit focus on teaching such skills. Future outcome research in subgroups who have relatively higher impairment in adaptive skills is needed. Additionally, the age-related differences in the adaptive behavior profiles among the IQ-matched children suggest a different trajectory of gains in adaptive skills in children with ASD as compared to typically developing children. This study analyzed the adaptive behavior profiles in two broad age groups. Further studies comparing smaller age groups and longitudinal outcomes are required to delineate the trajectory of adaptive behavior development in ASD. Furthermore, the current tools used to measure adaptive behavior rely on the standard scores based on the age-equivalent normative references obtained from a representative sample comprising of various clinical groups including autism, and it is not known if the adaptive behavior scores in ASD population can truly discriminate between a developmental deviation or delay. Additionally, VABS item subset scores were not included in the analysis, which may provide a more detailed assessment of specific adaptive behavior in ASD (Balboni et al., 2016). Further evaluation of trajectories of specific adaptive behavior development in ASD is vital to understand the significance of adaptive behavior scores in this population.

Study limitations

There are some important limitations that need to be acknowledged. First, the ATN data were collected at baseline visits from a clinical sample, and the study has inherent biases of an observational study, which may limit generalization of the results. Our sample had a heterogeneous demographic profile, where Caucasian children of highly educated parents were overrepresented. Second, although VABS was used in all children to measure adaptive profile, different tools were used to obtain overall IQ, as is common in ASD practice and research. In our cohort, overall IQ was reported from six different tests, and assuming comparability in IQ test scores is disputable. The majority of children in our cohort had a Stanford–Binet test for IQ assessment. We ran a sensitivity analysis including only these children, and the results did not change in significance and changed minimally in magnitude. Moreover, the IQ categories were defined by SD below the population mean and not by the descriptive labels used for the individual test. Hence, results for all the participants with available overall IQ are reported to make the results generalizable across various clinical settings. Using overall IQ score is another limitation, as children with autism have inherent differences between verbal and non-verbal IQ. We categorized the cohort in three groups; low (IQ < 70), borderline (IQ: 70–85), and average (IQ > 85), whereas most of the previous studies considered IQ of 70 as cut-off for low or high-functioning ASD. This proved to be advantageous, as the relative strengths and deficits of “borderline” intellectual functioning group are more apparent with this method although we recognize that IQ boundaries are no longer part of the classification of intellectual functioning in Diagnostic and Statistical Manual of Mental disorders (5th ed.; DSM-V; APA, 2013). Finally, these data were collected before DSM-5 was available and the results are based on DSM-IV-TR diagnosis. The diagnostic criteria for ASD have been modified, and this may have implication on receiving a diagnosis of ASD in current practice. Nevertheless, this study highlights the level of functional impairment that children with ASD experience and examines some interesting questions regarding the adaptive behavior profiles in children with ASD.

Supplemental Material

AUT733113_Lay_Abstract – Supplemental material for Correlates of adaptive behavior profiles in a large cohort of children with autism: The autism speaks Autism Treatment Network registry data

Supplemental material, AUT733113_Lay_Abstract for Correlates of adaptive behavior profiles in a large cohort of children with autism: The autism speaks Autism Treatment Network registry data by Manina Pathak, Amanda Bennett and Amy M Shui in Autism

Footnotes

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: This Network activity was supported by Autism Speaks and cooperative agreement UA3 MC11054 through the US Department of Health and Human Services, Health Resources and Services Administration, Maternal and Child Health Research Program to the Massachusetts General Hospital. This work was conducted through the Autism Speaks Autism Treatment Network.

ORCID iD

Amanda Bennett

References

American Psychiatric Association (APA) (2000) Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: APA.

American Psychiatric Association (APA) (2013) Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington: APA.

Bal

Kim

Cheong

et al . (2015) Daily living skills in individuals with autism spectrum disorder from 2 to 21 years of age. Autism 19: 774–784.

Balboni

Tasso

Muratori

et al . (2016) The Vineland-II in preschool children with autism spectrum disorders: an item content category analysis. Journal of Autism and Developmental Disorders 46: 42–52.

Bölte

Poustka

(2002) The relation between general cognitive level and adaptive behavior domains in individuals with autism with and without co-morbid mental retardation. Child Psychiatry & Human Development 33: 165–172.

Carter

Volkmar

Sparrow

et al . (1998) The Vineland Adaptive Behavior Scales: supplementary norms for individuals with autism. Journal of Autism and Developmental Disorders 28: 287–302.

Chakrabarti

Fombonne

(2005) Pervasive developmental disorders in preschool children: confirmation of high prevalence. American Journal of Psychiatry 162: 1133–1141.

Charman

Pickles

Simonoff

et al . (2011) IQ in children with autism spectrum disorders: data from the Special Needs and Autism Project (SNAP). Psychological Medicine 41: 619–627.

Constantino

Charman

(2016) Diagnosis of autism spectrum disorder: reconciling the syndrome, its diverse origins, and variation in expression. Lancet Neurology 15: 279–291.

10.

Dean

Harwood

Kasari

(2016) The art of camouflage: gender differences in the social behaviors of girls and boys with autism spectrum disorder. Autism 21: 678–689.

11.

Dworzynski

Ronald

Bolton

et al . (2012) How different are girls and boys above and below the diagnostic threshold for autism spectrum disorders? Journal of the American Academy of Child and Adolescent Psychiatry 51: 788–797.

12.

Eaves

(2008) Young adult outcome of autism spectrum disorders. Journal of Autism and Developmental Disorders 38: 739–747.

13.

Elliott

Murray

Pearson

(1990) Differential Ability Scales. San Antonio, TX: The Psychological Corporation.

14.

Farley

McMahon

Fombonne

et al . (2009) Twenty-year outcome for individuals with autism and average or near-average cognitive abilities. Autism Research 2: 109–118.

15.

Fenton

D’Ardia

Valente

et al . (2003) Vineland adaptive behavior profiles in children with autism and moderate to severe developmental delay. Autism 7: 269–287.

16.

Fombonne

(2009) Epidemiology of pervasive developmental disorders. Pediatric Research 65: 591–598.

17.

Frazier

Georgiades

Bishop

et al . (2014) Behavioral and cognitive characteristics of females and males with autism in the Simons Simplex Collection. Journal of the American Academy of Child and Adolescent Psychiatry 53: 329–340.

18.

Gabriels

Ivers

Hill

et al . (2007) Stability of adaptive behaviors in middle-school children with autism spectrum disorders. Research in Autism Spectrum Disorders 1: 291–303.

19.

Gillham

Carter

Volkmar

et al . (2000) Toward a developmental operational definition of autism. Journal of Autism and Developmental Disorders 30: 269–278.

20.

Howlin

(2013) Outcomes in Autism Spectrum Disorders. Hoboken, NJ: John Wiley & Sons, pp.201–220.

21.

Howlin

Goode

Hutton

et al . (2004) Adult outcome for children with autism. Journal of Child Psychology and Psychiatry 45: 212–229.

22.

Kanne

Gerber

Quirmbach

et al . (2011) The role of adaptive behavior in autism spectrum disorders: implications for functional outcome. Journal of Autism and Developmental Disorders 41: 1007–1018.

23.

Kenworthy

Case

Harms

et al . (2010) Adaptive behavior ratings correlate with symptomatology and IQ among individuals with high-functioning autism spectrum disorders. Journal of Autism and Developmental Disorders 40: 416–423.

24.

Klin

Saulnier

Sparrow

et al . (2007) Social and communication abilities and disabilities in higher functioning individuals with autism spectrum disorders: the Vineland and the ADOS. Journal of Autism and Developmental Disorders 37: 748–759.

25.

Kraijer

(2000) Review of adaptive behavior studies in mentally retarded persons with autism/pervasive developmental disorder. Journal of Autism and Developmental Disorders 30: 39–47.

26.

Liss

Harel

Fein

et al . (2001) Predictors and correlates of adaptive functioning in children with developmental disorders. Journal of Autism and Developmental Disorders 31: 219–230.

27.

Loomes

Hull

Mandy

WPL

(2017) What is the male-to-female ratio in autism spectrum disorder? A systematic review and meta-analysis. Journal of the American Academy of Child and Adolescent Psychiatry 56: 466–474.

28.

Lopata

Smith

Volker

et al . (2013) Comparison of adaptive behavior measures for children with HFASDs. Autism Research and Treatment 2013: 415989.

29.

Lord

Rutter

DiLavore

et al . (2012) ADOS-2: Autism Diagnostic Observation Schedule. 2nd ed. Torrance, CA: WPS.

30.

Matthews

Pollard

Ober-Reynolds

et al . (2015) Revisiting cognitive and adaptive functioning in children and adolescents with autism spectrum disorder. Journal of Autism and Developmental Disorders 45: 138–156.

31.

Mouga

Almeida

Cafe

et al . (2015) Adaptive profiles in autism and other neurodevelopmental disorders. Journal of Autism and Developmental Disorders 45: 1001–1012.

32.

Mullen

(1995) Mullen Scales of Early Learning Manual. Circle Pines, MN: American Guidance Service, Inc.

33.

Perry

Flanagan

Dunn Geier

et al . (2009) Brief report: the Vineland Adaptive Behavior Scales in young children with autism spectrum disorders at different cognitive levels. Journal of Autism and Developmental Disorders 39: 1066–1078.

34.

Pugliese

Anthony

Strang

et al . (2015) Increasing adaptive behavior skill deficits from childhood to adolescence in autism spectrum disorder: role of executive function. Journal of Autism and Developmental Disorders 45: 1579–1587.

35.

Roid

(2003) Stanford-Binet Intelligence Scales. 5th ed. Itasca, IL: Riverside Publishing.

36.

Schalock

Borthwick-Duffy

Bradley

et al . (2010) Intellectual Disability: Definition, Classification, and Systems of Supports. Washington, DC: American Association on Intellectual and Developmental Disabilities.

37.

Shattuck

Durkin

Maenner

et al . (2009) Timing of identification among children with an autism spectrum disorder: findings from a population-based surveillance study. Journal of the American Academy of Child and Adolescent Psychiatry 48: 474–483.

38.

Sparrow

Cicchetti

Balla

(2005) Vineland Adaptive Behavior Scales. 2nd ed. Circle Pines, MN: American Guidance Service.

39.

Tomanik

Pearson

Loveland

et al . (2007) Improving the reliability of autism diagnoses: examining the utility of adaptive behavior. Journal of Autism and Developmental Disorders 37: 921–928.

40.

Volkmar

Szatmari

Sparrow

(1993) Sex differences in pervasive developmental disorders. Journal of Autism and Developmental Disorders 23: 579–591.

41.

Wechsler

(1999) The Wechsler Abbreviated Scale of Intelligence. San Antonio, TX: The Psychological Corporation.

42.

Wechsler

(2002) WPPSI-III Administration and Scoring Manual. San Antonio, TX: The Psychological Corporation.

43.

Wechsler

(2003) Wechsler Intelligence Scale for Children. Administration and Scoring Manual. 4th ed. San Antonio, TX: Harcourt Assessment, Inc.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.12 MB