Skip to main content
NCBI home page
HHS Author Manuscripts logoLink to HHS Author Manuscripts
. Author manuscript; available in PMC: 2026 Mar 17.
Published before final editing as: J Autism Dev Disord. 2025 Aug 20:10.1007/s10803-025-06978-7. doi: 10.1007/s10803-025-06978-7

Associations of Adaptive Behavior and Wandering with Serious Injuries in Young Children with Autism: Study to Explore Early Development

C DiGuiseppi 1, B Holst 1, M Bolt 2, S J Schmiege 2, K Bartholomay 1, B Barger 3, T Crume 1, S Levy 4, C Nadler 5, L A Schieve 6, L H Tian 6, L D Wiggins 6
PMCID: PMC12990144  NIHMSID: NIHMS2151300  PMID: 40833658

Abstract

Purpose

Increased injury risk in children with autism may result from co-occurring developmental characteristics and challenges. We examined associations of adaptive behavior delay and wandering with injuries among children with autism.

Methods

Cross-sectional analyses were performed using data from 648 children aged 30–68 months with autism determined by standardized instruments. Associations of adaptive behavior delay and wandering with any parent-reported serious injury (i.e., resulting in emergency department visit or hospitalization) since birth were examined, adjusting for age and attention-deficit/hyperactivity problems. Effect modification by significant early learning delay (SELD) was assessed.

Results

Prevalence of serious injury was 27%. Delayed adaptive behavior occurred in 81% and wandering sometimes or often in 59%. Associations of both adaptive behavior and wandering with serious injury differed by SELD status. Among children with SELD, adjusted prevalence of serious injury was increased in those with normal adaptive behavior (47%) versus delayed adaptive behavior (24%) (adjusted prevalence ratio [aPR] = 2.01; 95% confidence interval (CI): 1.17, 3.45, p = 0.011) and in those who wandered often or sometimes (30%) versus those who did not (18%) (aPR = 1.50; 95%CI: 1.01, 2.25, p = 0.047). Among children without SELD, neither normal adaptive behavior (aPR = 0.82; 95%CI: 0.54, 1.32; p = 0.336) nor wandering (aPR = 0.88; 95%CI: 0.59, 1.30; p = 0.516) was significantly associated with serious injury.

Conclusion

Among children with autism, there appears to be a complex interplay among wandering and adaptive behavior, SELD and injury occurrence. Children with autism and SELD who have normal adaptive behavior or who wander may be important targets for injury prevention interventions.

Keywords: Autism spectrum disorder, Injuries, Wandering, Adaptive behavior

Introduction

Unintentional injuries are the leading cause of death among US children aged 3–5 years, accounting for 923 deaths and 56,462 years of potential life lost in 2022 (Centers for Disease Control and Prevention, 2024). Unintentional injuries also accounted for over 900,000 emergency department (ED) visits in this age group in 2022, resulting in 33,600 hospitalizations (Centers for Disease Control and Prevention, 2024). The estimated lifetime medical, work-loss, and quality of life costs from ED visits for nonfatal unintentional injuries to US children aged 3–5 years in 2022 that resulted in treatment and release or hospitalization totaled $73.0 billion (Centers for Disease Control and Prevention, 2024).

Autism and Injury Risk

Autism spectrum disorder (ASD), also known as autism (Lord et al., 2020), is a child developmental diagnosis defined by social communication and interaction deficits and restricted interests and repetitive behaviors. Children withautism1 are more likely to be male, display behavioral and cognitive problems, and have mothers with a history of psychiatric conditions (Daniels et al., 2008; Hodge et al., 2011; Lord et al., 2020; Lyall et al., 2017; Newschaffer et al., 2007; Yirmiya & Shaked, 2005), all of which are factors associated with increased risk for child injuries (Lyngsoe et al., 2021; Myhre et al., 2012; O’Donnell & Canares, 2021; Orton et al., 2012; Rowe et al., 2004, 2007; Ruiz-Goikoetxea et al., 2018; Schwebel & Gaines, 2007). Several studies have reported increased likelihood of injuries or injury deaths among individuals with autism compared to those without autism, after accounting for sociodemographic differences (Guan & Li, 2017; Lee et al., 2008; McDermott et al., 2008; Vohra et al., 2016). Other studies have reported a reduced risk of injury in children with autism after controlling for co-occurring conditions (e.g., attention-deficit/hyperactivity disorder [ADHD], anxiety, and depression) and learning delay or intellectual disability (Agnafors et al., 2020; Allan et al., 2021; DiGuiseppi et al., 2018). Both Allan et al. (2021), which examined selected non-accidental injuries, and DiGuiseppi et al. (2018), which examined all injuries, reported no increased odds of injury in preschool-age children with autism relative to population controls after accounting for factors such as attention problems and cognitive ability. Agnafors et al. (2020) reported no association of autism with injuries from falls, motor vehicle crashes, or other transport accidents in children aged 0–6 years after controlling for other psychiatric diagnoses (including ADHD, oppositional defiant disorder/conduct disorder, anxiety and affective disorder). These studies taken together suggest that excess injury risk reported in children with autism likely results at least in part from behavioral and emotional problems and cognitive delays that often co-occur with autism. Other behavioral factors may also contribute to injury risk in autism but have not been fully explored, in particular, delayed adaptive behavior and wandering.

Delayed Adaptive Behavior in Autism and Risk for Injury

Adaptive behavior reflects an individual’s social and practical competence to meet the demands of everyday living (Sparrow, et al., 2005). Individuals with autism can display deficits in adaptive functioning, both in core symptoms (i.e., functional socialization and communication skills) and more generally (e.g., daily living skills) (Saulnier et al., 2022). Difficulty navigating situations that might lead to injury, and errors in judging the relationship between one’s abilities and the demands of a situation, may contribute to injury risk (Schwebel & Barton, 2005). Several studies of typically developing children have suggested a link between young children’s overestimation of their abilities with increased risk of medically attended injuries or injury severity (Plumert, 1995; Plumert & Schwebel, 1997; Schwebel & Plumert, 1999), although similar studies in children with autism have not been conducted.

Wandering in Autism and Risk for Injury

Wandering, also termed elopement, is the tendency for an individual to leave the safety of a responsible person’s care or safe area (National Autism Association, 2017). Wandering is more common among children with autism than other children, regardless of age or time period assessed. Wiggins et al. (2020) reported that 60% of children ages 4–5 with autism wandered at least sometimes in the past 3 months, compared to 25% of children with other developmental delays and 12% of children from the general population. Rice et al. (2016) found that children ages 6–17 with autism with or without intellectual disability (ID) were about twice as likely to have wandered off in the past year as children with ID alone (43% vs. 24%) and Anderson et al., (2012) noted that 49% of children ages 4–17 with autism had ever eloped, whereas only 13% of their unaffected siblings had done so. Wandering may have an important impact on child injury risk. In a survey of families of children with autism, 26% reported that their child had wandered and was missing long enough to cause concern, of whom 24% were in danger of drowning and 65% in danger of traffic injury (Anderson et al., 2012). In an evaluation of 808 wandering events over a 5-year period in individuals with autism, identified through media and government agencies, 17% resulted in death (71% of these from drowning) and 13% required medical attention (National Autism Association, 2017). In studies of children with autism, wandering itself has been linked to behavioral and emotional problems such as ADHD, anxiety, depression, and conduct problems in some studies (McLaughlin et al., 2018; Rice et al., Wiggins et al., 2020) but not others (Anderson et al., 2012; Rice et al., 2016). Associations between wandering and cognitive delays among children with autism have also been reported (Anderson et al., 2012; Wiggins et al., 2020).

In this analysis of preschool children with autism from the Study to Explore Early Development (SEED), we examined the associations of delayed adaptive behavior and wandering with lifetime injury occurrence after accounting for other factors associated with injury risk, including behavioral problems and cognitive delay. We hypothesized that both delayed adaptive behavior and wandering would be associated with increased prevalence of serious injury in children with autism.

Methods

Study Design and Setting

SEED is a multi-site case-control study, for which methods have been previously detailed (Schendel et al., 2012). The six sites (California, Colorado, Georgia, Maryland, North Carolina, and Pennsylvania) that enrolled participants in SEED Phase 1 (SEED1, 2007–2011) were included in this analysis (as injury data were not collected in later SEED Phases). SEED1 was approved by institutional review boards at the Centers for Disease Control and Prevention and each study site.

Participants

Children were eligible for SEED1 enrollment if they were born between September 1, 2003, and August 31, 2006, in a study catchment area and lived in the same area at first study contact. The child also must have resided with a caregiver aged ≥ 18 years who had cared for the child continuously since the child was at least 6 months of age and who spoke English or, at two of the study sites, either English or Spanish. Demographics and characteristics of the different study catchment sites and the SEED study design and methods have been previously described (DiGuiseppi et al., 2016; Schendel et al., 2012). SEED participants were identified and recruited from specialized autism diagnostic and treatment centers, early intervention services providers, and vital records. Families were sent an introductory letter followed by a phone call to assess eligibility. Written informed consent was obtained from all participating families.

Procedures

Each parent or other caregiver (99% of whom were mothers) completed a telephone or in-person interview about family, child, and household characteristics, health conditions, and behaviors. Caregivers completed the Social Communication Questionnaire (SCQ) (Rutter et al., 2003) and their children received clinical developmental assessments when they were between 30.0 and 68.9 months old, including the Mullen Scales of Early Learning (MSEL), to assess earlylearning delays (Mullen, 1995). Children with a pre-existing diagnosis of autism, an SCQ score ≥ 11 (Allen et al., 2007; Lee et al., 2007) or observed signs of autism during MSEL administration were evaluated by trained clinicians using standardized instruments, including the Autism Diagnostic Observation Schedule (Lord, et al., 1999) and the Autism Diagnostic Interview-Revised (Gotham et al., 2007). Children were classified into the autism case group based on an algorithm developed by SEED clinicians (Schendel et al., 2012; Wiggins et al., 2015). Only the autism case group is included in the current cross-sectional analysis.

Primary Exposures

The Vineland Adaptive Behavior Scales, 2nd Edition (VABS-2) (Sparrow, et al., 2005), a reliable, validated standardized assessment tool that uses a semi-structured interview to assess adaptive behavior, and has been used extensively with heterogeneous groups of children with autism (Perry et al., 2009), was administered to caregivers. Composite standard scores between 71 and 85 suggest some difficulties in adaptive functioning while scores of 70 or below suggest significant difficulties. For this analysis, we dichotomized VABS-2 composite scores such that participants with scores ≤ 85 were categorized as “delayed” (i.e., at least some difficulties in adaptive functioning) while those with scores > 85 were categorized as “not delayed.” Caregivers completed the Child Behavior Checklist (CBCL) for ages 1.5–5 years, which has been shown to have high testretest and interrater reliability and to discriminate between children who were or were not referred for mental health or special education services (Achenbach & Rescorla, 2000). The CBCL is a standardized instrument containing 99 behaviors that are rated as “not true,” “somewhat or sometimes true,” or “very true or often true” within the past 6 months. One item asks whether “wanders away” describes the child; the term “wanders away” is not defined. The term “elopement” is not included in the CBCL. Responses to this item, categorized as sometimes/often true versus not true according to parent report, formed the “wandering” exposure.

Outcome

A structured telephone interview with caregivers included questions about whether the child had experienced any injuries since birth that needed medical attention and, for each such injury, whether it had resulted in an ED visit or hospitalization. In addition, a free text description (“what was the injury?”) was collected. Three investigators independently categorized each free text injury description; parent-reported ‘injuries’ that were agreed by the investigators not to represent an acute physical injury were excluded. The outcome for this analysis was at least one “serious injury,” which we define as any injury resulting in an ED visit or hospitalization (recognizing that such injuries may vary in their severity). We chose this outcome because maternal recall of medically-attended injuries among their children aged < 6 years is more accurate for injuries resulting in an ED visit or hospitalization than for injuries that are seen in other settings (Cummings et al., 2005). Descriptions of the frequency, nature, anatomic region, and external circumstances of injuries in the study sample have been reported previously (DiGuiseppi et al., 2018).

Other Variables

Behavioral risk factors were assessed using Diagnostic and Statistical Manual of Mental Disorders (DSM)-oriented subscales derived from the CBCL, specifically, the Anxiety Problems, Attention-Deficit/Hyperactivity Problems and Oppositional Defiant Problems subscales, with t-scores ≥ 65 indicating borderline or clinically significant problems in the child. We used DSM-oriented subscales because the CBCL wandering item defining our primary exposure is not included in the items used to derive these subscales. Early learning was assessed using the MSEL Early Learning Composite (ELC) score, with scores < 70 defining ‘significant early learning delay’ (SELD). Sociodemographic factors assessed included child sex at birth and age at enrollment; maternal education, race/ethnicity, language spoken at home and age at child’s birth; and ratio of household income to federal poverty level. Maternal history of ever having physician-diagnosed depression was also assessed (yes/no).

Statistical Analysis

Descriptive statistics were calculated using means, standard deviations, and ranges for continuous data and counts and percentages for categorical data. Associations between each of the two exposures of interest (i.e., delayed adaptive behavior and wandering) and the injury outcome were examined using separate multivariable robust Poisson models to estimate prevalence ratios (PRs) and 95% confidence intervals (CIs). Based on studies of injury risk (Amiri et al., 2017; Myhre et al., 2012; O’Donnell & Canares, 2021; Rowe et al., 2004; Ruiz-Goikoetxea et al., 2018; Schwebel & Gaines, 2007), wandering (Anderson et al., 2012; McLaughlin et al., 2018; Wiggins et al., 2020) and delayed adaptive behavior (Saulnier et al., 2022), we included attention-deficit/hyperactivity problems (categorized as clinical/borderline vs. normal), SELD (yes vs. no), and child age at enrollment as confounding variables in all multivariable models. Interactions on a multiplicative scale were evaluated between each of the two primary exposure variables and both attention-deficit/hyperactivity problems and SELD using a regression framework; interaction terms that were statistically significant at p < 0.05 in either unadjusted model were retained in both models. In post-hoc analyses, among children with SELD, we calculated mean differences in MSEL ELC scores between those who did and did not have delayed adaptive behavior. We performed similar calculations for children with SELD between those who did and did not wander. Available-data analysis (resulting in differing datasets for analyses of delayed adaptive behavior and wandering) was used to manage missing data. R version 4.3.1 (2023-06-16)(R Core Team, 2023) and the packages tidyverse (v2.0.0) (Wickham et al., 2019), sjPlot (v2.8.15) (Lüdecke, 2023), sandwich (v3.0.2) (Zeileis, 2006; Zeileis et al.,2020), and emmeans (v1.10.3) (Lenth, 2024) were used for data manipulation and analysis.

Results

Sample Characteristics

Of 693 children classified with autism whose mothers completed the history of postnatal injuries (DiGuiseppi et al., 2018), 648 children (94%) had CBCL data, of whom 640 (92%) with MSEL data were included in the analytic cohort for evaluation of wandering. The analytic cohort for evaluation of delayed adaptive behavior was 628 (90%) after excluding 12 children with missing data for the VABS-2. Within these two analytic cohorts, there were no missing data for the covariates included in the final models.

Among 648 children in the sample, 173 (27%) had at least one serious injury. Mean child age in the sample was 55.8 (SD 6.8) months and 82% were male. Substantial percentages had clinical/borderline scores on the AttentionDeficit/Hyperactivity, Anxiety and/or Oppositional Defiant Problems subscales (29%, 19%, and 25%, respectively). The majority (62%) had SELD. Most mothers were aged 25 years or older at child’s birth (87%), were non-Hispanic white race (56%), spoke English at home (88%), had a bachelor’s or advanced degree (52%) and reported current household income that was less than 400% of the Federal Poverty Level (65%). Compared to children who did not have a serious injury, children who had ever had a serious injury were older and had a higher prevalence of clinical/borderline attention-deficit/hyperactivity problems. The presence of serious injury also differed by maternal education and language spoken at home (Table 1).

Table 1.

Characteristics of children with autism aged 3–5 years who ever had an injury resulting in an emergency department visit or hospitalization (“serious injury”) versus never did, Study to Explore Early Development, 2007–2011

No serious injury
(N=475)
N (%)/Mean (SD)		 Any serious injury
(N = 173)
N (%)/Mean (SD)		 p-value
Sociodemographic & Maternal Characteristics
Child age (months) 55.4 (7.0) 57.0 (6.3) 0.007
Child sex male 389 (81.9%) 141 (81.5%) 0.909
Maternal age at child’s birth (years) 31.3 (5.6) 31.0 (5.6) 0.546
Maternal ethnicity/race 0.495
Non-Hispanic White 257 (54.2%) 104 (60.1%)
Non-Hispanic Black 100 (21.1%) 28 (16.2%)
Non-Hispanic Othera 58 (12.2%) 20 (11.6%)
Hispanic, any race 59 (12.4%) 21 (12.1%)
Language spoken at home 0.011
English 406 (85.8%) 163 (94.2%)
Spanish 31 (6.6%) 3 (1.7%)
Other 36 (7.6%) 7 (4.0)
Maternal education at enrollment 0.030
High school degree or less 105 (22.1%) 25 (14.5%)
Some college 124 (26.1%) 60 (34.7%)
Bachelor’s degree or higher 246 (51.8%) 88 (50.9%)
Family Income as % of federal poverty level 0.601
<139% 122 (26.8%) 37 (21.8%)
139 - <=250% 65 (14.3%) 28 (16.5%)
251 - <=399% 111 (24.3%) 45 (26.5%)
>399% 158 (34.6%) 60 (35.3%)
Maternal history of depression 0.250
Yes 117 (25.3%) 49 (29.9%)
No 346 (74.7%) 115 (70.1%)
Child Behavior and Development
Adaptive Behavior Delay 0.564
Yes 379 (81.3%) 134 (79.3%)
No 87 (18.7%) 35 (20.7%)
Wandering 0.173
Sometimes/Often 271 (57.1%) 109 (63.0%)
None 204 (42.9%) 64 (37.0%)
Early Learning Delay 0.311
Yes 298 (63.7%) 102 (59.3%)
No 170 (36.3%) 70 (40.7%)
Attention-Deficit/Hyperactivity Problems 0.032
Yes 129 (27.2%) 62 (35.8%)
No 346 (72.8%) 111 (64.2%)
Anxiety Problems 0.151
Yes 95 (20.0%) 26 (15.0%)
No 380 (80.0%) 147 (85.0%)
Oppositional Defiant Problems 0.303
Yes 113 (23.8%) 48 (27.7%)
No 362 (76.2%) 125 (72.3%)
Open in a new tab

Missing data: race/ethnicity n = 1 (0.2%), language spoken at home n = 2 (0.3%), family income n = 22 (3.4%), maternal history of depression n = 21 (3.2%), adaptive behavior delay n = 13 (2.0%), significant early learning delay n = 8 (1.2%)

a

Includes American Indian or Alaska Native, Asian, Native Hawaiian or Other Pacific Islander, or multiple races.

Delayed Adaptive Behavior and Serious Injury

Delayed adaptive behavior was reported for 81% of sample children and was defined as the reference group. The prevalence of at least one serious injury was 29% in children without adaptive behavior delay and 26% in children with adaptive behavior delay. In unadjusted analysis, the prevalence of having been seriously injured was similar in those who did not have delayed adaptive behavior compared to those who did (PR = 1.10; 95% CI: 0.80, 1.51; p = 0.56). The association did not differ significantly based on the presence of attention-deficit/hyperactivity problems (pinteraction=0.071) but did differ based on the presence of SELD (pinteraction=0.047); therefore, the interaction term for SELD was retained in the adjusted model. Among children with SELD, the prevalence of injuries was almost twice as high in those without adaptive behavior delay (47%) versus with delay (24%), whereas among children without SELD, prevalence was 26% in those without adaptive behavior delay versus 32% in those with delay (Fig. 1).

Fig. 1.

Fig. 1

Open in a new tab

Prevalence of injuries resulting in an emergency department (ED) visit or hospitalization among children with and without delayed adaptive behavior (A) and among children who wander sometimes/often versus none (B), according to presence or absence of significant early learning delay, Study to Explore Early Development, 2007–2011

Because of the significant interaction between SELD and adaptive behavior, adjusted prevalence ratios are presented at each level of SELD. The interaction term remained statistically significant in the adjusted model (pinteraction=0.009). Among children with SELD, the adjusted prevalence of serious injury was twice as high among those without delayed adaptive behavior as those with delay (Table 2).

Table 2.

Association between adaptive behavior and ever having an injury resulting in an emergency department visit or hospitalization among children with autism, Study to Explore Early Development, 2007–2011

Any injury resulting in emergency department visit or hospitalization
Adjusted Prevalence Ratioa (95% CI) p-value
Adaptive behavior not delayed (vs. delayed) among children with significant early learning delayb 2.01 (1.17, 3.45) 0.011
Adaptive behavior not delayed (vs. delayed) among children without significant early learning delayb 0.82 (0.54, 1.32) 0.336
Open in a new tab
a

N = 628 children with autism. Model terms include child age in months, presence/absence of attention-deficit/hyperactivity problems, presence/absence of significant early learning delay, and adaptive behavior delay x significant early learning delay interaction

b

Coefficients for “adaptive behavior not delayed” interpreted at each level of early learning delay because of interaction of adaptive behavior delay with significant early learning delay (pinteraction=0.009 in adjusted model)

In post-hoc analysis, we observed that among children with SELD, MSEL ELC scores were significantly higher in children without delayed adaptive behavior versus with delay (mean 60.4 [5.8] vs. 53.3 [6.2]), mean difference [MD] = 7.1, 95%CI: 4.08, 10.16, p < 0.001). Among children without SELD, adaptive behavior was not associated with serious injury in the adjusted model (Table 2).

Wandering and Serious Injury

Wandering often or sometimes was reported for 59% of children in the sample. The prevalence of at least one serious injury was 29% in children who wandered sometimes or often and 24% in children who did not wander. In unadjusted analysis, wandering often or sometimes was not significantly associated with prevalence of having been injured (PR = 1.20; 95% CI 0.92, 1.57; p = 0.18). The association did not differ based on the presence of attention-deficit/hyperactivity problems (pinteraction=0.34). The interaction between wandering and SELD was not statistically significant (pinteraction=0.088), but the interaction term was retained in the adjusted model following a similar model structure as for delayed adaptive behavior. Among children with SELD, the prevalence of injuries was 30% in children who wandered often or sometimes and 18% in children who did not wander, whereasamong children without SELD, prevalence was 30% and 28%, respectively (Fig. 1).

Adjusted prevalence ratios are presented at each level of SELD. The interaction term was borderline statistically significant in the adjusted model (pinteraction=0.058). Among children with SELD, those who wandered had 1.5 times higher prevalence of ED visit or hospitalization for injury than those who did not wander in the adjusted model (Table 3).

Table 3.

Association between wandering and ever having an injury resulting in an emergency department visit or hospitalization among children with autism, Study to Explore Early Development, 2007–2011

Any injury resulting in emergency department visit or hospitalization
Adjusted Prevalence Ratioa (95% CI) p-value
Wanders sometimes/often (vs. none) among children with significant early learning delayb 1.50 (1.01–2.25) 0.047
Wanders sometimes/often (vs. none) among children without significant early learning delayb 0.88 (0.59–1.30) 0.516
Open in a new tab
a

N = 640 children with autism. Model terms include child age in months, presence/absence of attention-deficit/hyperactivity problems, presence/absence of significant early learning delay, and wandering x significant early learning delay interaction

b

Coefficients for “wanders sometimes/often” interpreted at each level of significant early learning delay because of interaction of wandering with significant early learning delay (pinteraction=0.058 in adjusted model)

In post-hoc analysis, we observed that within the SELD stratum, MSEL ELC scores were slightly lower in children who wandered than in those who did not (mean 53.0 [SD 5.9] vs. 54.8 [6.9], MD= −1.8 [95% CI: −3.16, −0.49], p = 0.008). Among children without SELD, wandering was not associated with injury in the adjusted model (Table 3).

Discussion

In previous analyses of children enrolled in the SEED study, we found that autism was not associated with serious injury after accounting for cognitive ability and attention problems (DiGuiseppi et al., 2018). The current analysis extends these findings to assess whether either delayed adaptive behavior or wandering behavior is associated with serious injury in children with autism after accounting for these same conditions. Results suggest that there are complex relationships of both adaptive behavior and wandering with serious injuries that are modified by SELD. Among children with autism without SELD, we found that neither adaptive behavior nor wandering was associated with serious injuries after accounting for child age and attention-deficit/hyperactivity problems. However, among children with autism who had SELD, we found both adaptive behavior and wandering were associated with serious injuries after accounting for these same variables. Considering child cognitive abilities along with adaptive behavior and wandering may improve development of injury prevention plans for young children with autism.

In terms of adaptive behavior, children with autism and SELD who did not have delayed adaptive behavior had an increased prevalence of serious injuries relative to those who had delayed adaptive behavior. This was contrary to the hypothesized relationship. One potential explanation is that children with autism who have dual delays may need and therefore receive more services and supports that involve close supervision, thereby potentially reducing injury risk. In post-hoc analyses, we observed that among children with SELD, those with delayed adaptive behavior had significantly lower ELC scores than those without adaptive behavior deficits, indicating substantially greater impairment. Another possible explanation for these findings is parental overestimation of child skills relevant to safety when there are discrepancies in cognitive and adaptive behavior functioning. Previous studies suggest that parents of young children in general tend to overestimate their children’s safety awareness and skills (Dunne et al., 1992; Miller & Davis, 1992; Miller et al., 1991; Morrongiello & Barton, 2009).

One study of children with autism reported that parents were more likely to overestimate adaptive abilities– which include safety awareness– relative to teacher assessments when cognitive abilities were lower (Dickson et al., 2018). Parents of children with autism and SELD but normal adaptive functioning could misjudge how well their child can assess and respond to dangerous situations, placing them at increased risk of injury. If so, parents of children with autism who have SELD with normal adaptive functioning may be an important target for parental safety counseling and other injury prevention interventions.

As far as wandering, children with autism and SELD who wandered had an increased prevalence of serious injury relative to those who did not wander, even after accounting for age and attention-deficit/hyperactivity problems. As noted above, some children with autism and SELD may receive additional supports, such as behavioral aides, that provide additional supervision and reduce risk for both wandering and serious injuries. For example, parents of children with autism rated the overall effectiveness of having individual behavioral aides or behavioral specialists as good or very good for preventing wandering (Anderson, et al., 2012). Children with autism and SELD who do not have these types of supports may benefit from counseling about environmental modifications that may reduce the risks of both wandering and serious injuries. For example, electronic tracking devices have been associated with both decreased frequency and duration of wandering and decreased risk of serious injury because of wandering (McLaughlin et al., 2020). Other strategies widely recommended to reduce child injuries, e.g., pool fencing, window guards/locks, safety gates, and constant adult supervision around water (Gardner et al., 2007; Hagan et al., 2017), may also prevent or reduce wandering. We were unable to assess this directly as we did not collect data on prevention strategies. Regardless, interventions designed to prevent wandering may potentially prevent injuries and vice versa.

Strengths and Limitations

There were potential limitations to this study. The primary outcome was maternal recall of any medically attended injury that resulted in an ED visit or hospitalization in the child’s lifetime (i.e., from birth to 3–5 years of age, depending on age at enrollment). Maternal recall of medically-attended injuries among their children aged < 6 years declines with time, although accuracy is better for injuries resulting in an ED visit or hospitalization (Cummings et al., 2005), as in our study. Nevertheless, if recall of injuries were differential depending on the child’s adaptive behavior, cognitive functioning or behavior problems, measures of association may have been biased. In addition, some serious injuries may be attended in other clinical settings, such as urgent care settings, which may have led to underestimates of the primary outcome. SEED’s relatively low recruitment contact rate (Schendel et al., 2012) may have increased the potential for biased measures of association if families with certain sociodemographic characteristics responded disproportionately to study invitations. However, we did not find evidence that sociodemographic differences (other than child age) confounded our findings. This study also has several strengths, including use of research-reliable administration of standardized instruments to evaluate and classify children with autism, inclusion of children who had not previously received a medical diagnosis of autism (perhaps reflecting less severe symptoms or lack of access to care), and collection of comprehensive data that enabled us to examine important covariates known to be associated with injury risk (Schendel et al., 2012).

Conclusion

Among children with autism, there appears to be a complex interplay among wandering and adaptive behavior and the occurrence of injuries serious enough to result in an ED visit or hospitalization that is modified by presence of SELD. In children with SELD, wandering is an independent predictor of serious injury. Moreover, children with SELD without corresponding adaptive deficits are at increased risk of serious injury compared to those with dual delays. These findings can inform healthcare providers who counsel parents of children with autism on factors associated with serious injury and strategies to reduce injury. They can also be used by parents and healthcare providers to develop and implement appropriate injury prevention strategies for this population.

Acknowledgements

The investigators acknowledge the contributions made to this study by project staff and enrolled families. Preliminary results from this study were presented at the International Meeting for Autism Research, Austin, TX, May 13, 2022.

Funding

The study was supported by six cooperative agreements from the Centers for Disease Control and Prevention (CDC): Cooperative Agreement Number U10DD000180, Colorado Department of Public Health/University of Colorado School of Medicine; Cooperative Agreement Number U10DD000181, Kaiser Foundation Research Institute (CA); Cooperative Agreement Number U10DD000182, University of Pennsylvania; Cooperative Agreement Number U10DD000183, Johns Hopkins University; Cooperative Agreement Number U10DD000184, University of North Carolina at Chapel Hill; and Cooperative Agreement Number U10DD000498, Michigan State University. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC.

Footnotes

Competing Interests The authors have no relevant competing interests to declare.

1

The authors acknowledge that differing opinions exist about the use of person-first (children with autism) versus identify-first (autistic children) language. For consistency with cited literature, we have used person-first language.

References

  1. Achenbach TM, & Rescorla LA (2000). Manual for the ASEBA preschool forms & profiles: An integrated system of multi-informant assessment; child behavior checklist for ages 1 1/2–5; Language development survey; caregiver teacher report form. University of Vermont. [Google Scholar]
  2. Agnafors S, Torgerson J, Rusner M, & Kjellstrom AN (2020). Injuries in children and adolescents with psychiatric disorders. Bmc Public Health, 20(1), 1273. 10.1186/s12889020092833 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Allan CC, DeShazer M, Staggs VS, Nadler C, Crawford TP, Moody S, & Chacko A (2021). Accidental injuries in preschoolers: Are we missing an opportunity for early assessment and intervention?? Journal of Pediatric Psychology, 46(7), 835–843. 10.1093/jpepsy/jsab044 [DOI] [PubMed] [Google Scholar]
  4. Allen CW, Silove N, Williams K, & Hutchins P (2007). Validity of the social communication questionnaire in assessing risk of autism in preschool children with developmental problems. Journal of Autism and Developmental Disorders, 37(7), 1272–1278. 10.1007/s1080300602797 [DOI] [PubMed] [Google Scholar]
  5. Amiri S, Sadeghi-Bazargani H, Nazari S, Ranjbar F, & Abdi S (2017). Attention deficit/hyperactivity disorder and risk of injuries: A systematic review and meta-analysis. Journal of Injury & Violence Research, 9(2), 95–105. 10.5249/jivr.v9i2.858 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Anderson C, Law JK, Daniels A, Rice C, Mandell DS, Hagopian L, & Law PA (2012). Occurrence and family impact of elopement in children with autism spectrum disorders. Pediatrics, 130(5), 870–877. 10.1542/peds.20120762 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Centers for Disease Control and Prevention (2024). Injury Prevention & Control: Data & Statistics (WISQARS™) (http://www.cdc.gov/injury/wisqars
  8. Cummings P, Rivara FP, Thompson RS, & Reid RJ (2005). Ability of parents to recall the injuries of their young children. Injury Prevention, 11(1), 43–47. 10.1136/ip.2004.006833 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Daniels JL, Forssen U, Hultman CM, Cnattingius S, Savitz DA, Feychting M, & Sparen P (2008). Parental psychiatric disorders associated with autism spectrum disorders in the offspring. Pediatrics, 121(5), e1357–1362. 10.1542/peds.20072296 [DOI] [PubMed] [Google Scholar]
  10. Dickson KS, Suhrheinrich J, Rieth SR, & Stahmer AC (2018). Parent and teacher concordance of child outcomes for youth with autism spectrum disorder. Journal of Autism and Developmental Disorders, 48(5), 1423–1435. 10.1007/s10803-017-3382-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. DiGuiseppi CG, Daniels JL, Fallin DM, Rosenberg SA, Schieve LA, Thomas KC, Windham GC, Goss CW, Soke GN, Currie DW, Singer AB, Lee LC, Bernal P, Croen LA, Miller LA, Pinto-Martin JA, Young LM, & Schendel DE (2016). Demographic profile of families and children in the study to explore early development (SEED): Casecontrol study of autism spectrum disorder. Disability and Health Journal, 9(3), 544–551. 10.1016/j.dhjo.2016.01.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. DiGuiseppi C, Levy SE, Sabourin KR, Soke GN, Rosenberg S, Lee LC, Moody E, & Schieve LA (2018). Injuries in children with autism spectrum disorder: Study to explore early development (SEED). Journal of Autism and Developmental Disorders, 48(2), 461–472. 10.1007/s10803-017-3337-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dunne RG, Asher KN, & Rivara FP (1992). Behavior and parental expectations of child pedestrians. Pediatrics, 89(3), 486–490. 10.1542/peds.89.3.486 [DOI] [PubMed] [Google Scholar]
  14. Gardner HG, American Academy of Pediatrics Committee on Injury, V., &, & Poison P (2007). Office-based counseling for unintentional injury prevention. Pediatrics, 119(1), 202–206. 10.1542/peds.2006-2899 [DOI] [PubMed] [Google Scholar]
  15. Gotham K, Risi S, Pickles A, & Lord C (2007). The autism diagnostic observation schedule: Revised algorithms for improved diagnostic validity. Journal of Autism and Developmental Disorders, 37(4), 613–627. 10.1007/s10803-006-0280-1 [DOI] [PubMed] [Google Scholar]
  16. Guan J, & Li G (2017). Injury mortality in individuals with autism. American Journal of Public Health. 10.2105/AJPH.2017.303696 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hagan JF, Shaw JS, & Duncan PM (2017). Bright futures: guidelines for health supervision of infants, children, and adolescents (Fourth edition. ed.). Bright Futures/American Academy of Pediatrics. [Google Scholar]
  18. Hodge D, Hoffman CD, & Sweeney DP (2011). Increased psychopathology in parents of children with autism: Genetic liability or burden of caregiving?? Journal of Developmental and Physical Disabilities, 23(3), 227–239. 10.1007/s10882-010-9218-9 [DOI] [Google Scholar]
  19. Lee LC, David AB, Rusyniak J, Landa R, & Newschaffer CJ (2007). Performance of the social communication questionnaire in children receiving preschool special education services. Research in Autism Spectrum Disorders, 1(2), 126–138. 10.1016/j.rasd.2006.08.004 [DOI] [Google Scholar]
  20. Lee LC, Harrington RA, Chang JJ, & Connors SL (2008). Increased risk of injury in children with developmental disabilities. Research in Developmental Disabilities, 29(3), 247–255. 10.1016/j.ridd.2007.05.002 [DOI] [PubMed] [Google Scholar]
  21. Lenth R (2024). emmeans: Estimated Marginal Means, aka Least- Squares Means. R package version 1.10.3. In https://CRAN.R-project.org/package=emmeans
  22. Lord C, Rutter M, DiLavore PC, & Risi S (1999). Autism diagnostic observation Schedule-WPS (ADOS-WPS). Western Psychological Services. [Google Scholar]
  23. Lord C, Brugha TS, Charman T, Cusack J, Dumas G, Frazier T, Jones EJH, Jones RM, Pickles A, State MW, Taylor JL, & Veenstra-VanderWeele J (2020). Autism spectrum disorder. Nature Reviews Disease Primers, 6(1), 5. 10.1038/s41572-019-0138-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lüdecke D (2023). sjPlot: Data Visualization for Statistics in Social Science. R package version 2.8.15. In https://CRAN.R-project.org/package=sjPlot
  25. Lyall K, Croen L, Daniels J, Fallin MD, Ladd-Acosta C, Lee BK, Park BY, Snyder NW, Schendel D, Volk H, Windham GC, & Newschaffer C (2017). The changing epidemiology of autism spectrum disorders. Annual Review of Public Health, 38, 81–102. 10.1146/annurev-publhealth-031816-044318 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lyngsoe BK, Munk-Olsen T, Vestergaard CH, Rytter D, Christensen KS, & Bech BH (2021). Maternal depression and childhood injury risk: A population-based cohort study in Denmark. Brain and Behavior, 11(3), e02029. 10.1002/brb3.2029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. McDermott S, Zhou L, & Mann J (2008). Injury treatment among children with autism or pervasive developmental disorder. Journal of Autism and Developmental Disorders, 38(4), 626–633. 10.1007/s10803-007-0426-9 [DOI] [PubMed] [Google Scholar]
  28. McLaughlin L, Keim SA, & Adesman A (2018). Wandering by children with autism spectrum disorder: Key clinical factors and the role of schools and pediatricians. Journal of Developmental and Behavioral Pediatrics, 39(7), 538–546. 10.1097/DBP.0000000000000591 [DOI] [PubMed] [Google Scholar]
  29. McLaughlin L, Rapoport E, Keim SA, & Adesman A (2020). Wandering by children with autism spectrum disorders: Impact of electronic tracking devices on elopement behavior and quality of life. Journal of Developmental and Behavioral Pediatrics, 41(7), 513–521. 10.1097/DBP.0000000000000817 [DOI] [PubMed] [Google Scholar]
  30. Miller SA, & Davis TL (1992). Beliefs about children: A comparative study of mothers, teachers, peers, and self. Child Development, 63(5), 1251–1265. 10.1111/j.1467-8624.1992.tb01693.x [DOI] [Google Scholar]
  31. Miller SA, Manhal M, & Mee LL (1991). Parental beliefs, parental accuracy, and children’s cognitive performance. Developmental Psychology, 27(2), 267–276. 10.1037/0012-1649.27.2.267 [DOI] [Google Scholar]
  32. Morrongiello BA, & Barton BK (2009). Child pedestrian safety: Parental supervision, modeling behaviors, and beliefs about child pedestrian competence. Accident Analysis and Prevention, 41(5), 1040–1046. 10.1016/j.aap.2009.06.017 [DOI] [PubMed] [Google Scholar]
  33. Mullen E (1995). Mullen scales of early learning. American Guidance Service, Inc. [Google Scholar]
  34. Myhre MC, Thoresen S, Grøgaard JB, & Dyb G (2012). Familial factors and child characteristics as predictors of injuries in toddlers: A prospective cohort study. British Medical Journal Open, 2, e000740. 10.1136/bmjopen-2011-000740 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. National Autism Association (2017). Mortality & Risk In ASD Wandering/Elopement 2011–2016. https://nationalautismassociation.org/wp-content/uploads/2017/04/NAAMortalityRiskASDElopement.pdf
  36. Newschaffer CJ, Croen LA, Daniels J, Giarelli E, Grether JK, Levy SE, Mandell DS, Miller LA, Pinto-Martin J, Reaven J, Reynolds AM, Rice CE, Schendel D, & Windham GC (2007). The epidemiology of autism spectrum disorders. Annual Review of Public Health, 28, 235–258. 10.1146/annurev.publhealth.28.021406.144007 [DOI] [PubMed] [Google Scholar]
  37. O’Donnell EP, & Canares TL (2021). Accidents waiting to happen: A review of unintentional household injuries in children. Pediatrics in Review, 42(3), 109–122. 10.1542/pir.2019-0169 [DOI] [PubMed] [Google Scholar]
  38. Orton E, Kendrick D, West J, & Tata LJ (2012). Independent risk factors for injury in pre-school children: Three population-based nested case-control studies using routine primary care data. PLOS One, 7(4), e35193. 10.1371/journal.pone.0035193 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Perry A, Flanagan HE, Geier D,J, & Freeman NL (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(7), 1066–1078. 10.1007/s10803-009-0704-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Plumert JM (1995). Relations between children’s overestimation of their physical abilities and accident proneness. Developmental Psychology, 31(5), 866–876. 10.1037/0012-1649.31.5.866 [DOI] [Google Scholar]
  41. Plumert JM, & Schwebel DC (1997). Social and temperamental influences on children’s overestimation of their physical abilities: Links to accidental injuries. Journal of Experimental Child Psychology, 67(3), 317–337. 10.1006/jecp.1997.2411 [DOI] [PubMed] [Google Scholar]
  42. R Core Team (2023). R: A language and environment for statistical computing. In R Foundation for Statistical Computing. https://www.R-project.org/ [Google Scholar]
  43. Rice CE, Zablotsky B, Avila RM, Colpe LJ, Schieve LA, Pringle B, & Blumberg SJ (2016). Reported wandering behavior among children with autism spectrum disorder and/or intellectual disability. Journal of Pediatrics, 174, 232–239e232. 10.1016/j.jpeds.2016.03.047 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Rowe R, Maughan B, & Goodman R (2004). Childhood psychiatric disorder and unintentional injury: Findings from a National cohort study. Journal of Pediatric Psychology, 29(2), 119–130. 10.1093/jpepsy/jsh015 [DOI] [PubMed] [Google Scholar]
  45. Rowe R, Simonoff E, & Silberg JL (2007). Psychopathology, temperament and unintentional injury: Cross-sectional and longitudinal relationships. Journal of Child Psychology and Psychiatry and Allied Disciplines, 48(1), 71–79. 10.1111/j.1469-7610.2006.01674.x [DOI] [PubMed] [Google Scholar]
  46. Ruiz-Goikoetxea M, Cortese S, Aznarez-Sanado M, Magallon S, Alvarez Zallo N, Luis EO, de Castro-Manglano P, Soutullo C, & Arrondo G (2018). Risk of unintentional injuries in children and adolescents with ADHD and the impact of ADHD medications: A systematic review and meta-analysis. Neuroscience and Biobehavioral Reviews, 84, 63–71. 10.1016/j.neubiorev.2017.11.007 [DOI] [PubMed] [Google Scholar]
  47. Rutter ML, Bailey A, & Lord C (2003). Social Communication Questionnaire (SCQ): Manual. [Google Scholar]
  48. Saulnier CA, Klaiman C, & McQueen E (2022). Adaptive behavior profiles in autism spectrum disorder. Current Psychiatry Reports, 24(12), 749–756. 10.1007/s11920-022-01381-w [DOI] [PubMed] [Google Scholar]
  49. Schendel DE, DiGuiseppi C, Croen LA, Fallin MD, Reed PL, Schieve LA, Wiggins LD, Daniels J, Grether J, Levy SE, Miller L, Newschaffer C, Pinto-Martin J, Robinson C, Windham GC, Alexander A, Aylsworth AS, Bernal P, Bonner JD, & Yeargin-Allsopp M (2012). The study to explore early development (SEED): A multisite epidemiologic study of autism by the centers for autism and developmental disabilities research and epidemiology (CADDRE) network. Journal of Autism and Developmental Disorders, 42(10), 2121–2140. 10.1007/s10803-012-1461-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Schwebel DC, & Barton BK (2005). Contributions of multiple risk factors to child injury. Journal of Pediatric Psychology, 30(7), 553–561. 10.1093/jpepsy/jsi042 [DOI] [PubMed] [Google Scholar]
  51. Schwebel DC, & Gaines J (2007). Pediatric unintentional injury: Behavioral risk factors and implications for prevention. Journal of Developmental and Behavioral Pediatrics, 28(3), 245–254. 10.1097/01.DBP.0000268561.80204.2a [DOI] [PubMed] [Google Scholar]
  52. Schwebel DC, & Plumert JM (1999). Longitudinal and concurrent relations among temperament, ability estimation, and injury proneness. Child Development, 70(3), 700–712. 10.1111/1467-8624.00050 [DOI] [PubMed] [Google Scholar]
  53. Sparrow SS, Cicchetti D, & Balla DA (2005). Vineland Adaptive Behavior Scales, Second Edition. In: APA PsycTests. [Google Scholar]
  54. Vohra R, Madhavan S, & Sambamoorthi U (2016). Emergency department use among adults with autism spectrum disorders (ASD). Journal of Autism and Developmental Disorders, 46(4), 1441–1454. 10.1007/s10803-015-2692-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Wickham H, Averick M, Bryan J, Chang W, McGowan L, François R, Grolemund G, Hayes A, Henry L, Hester J, Kuhn M, Pedersen T, Miller E, Bache S, Müller K, Ooms J, Robinson D, Seidel D, Spinu V, & Yutani H (2019). Welcome to the tidyverse. Journal of Open Source Software, 4, 1686. 10.21105/joss.01686 [DOI] [Google Scholar]
  56. Wiggins LD, Reynolds A, Rice CE, Moody EJ, Bernal P, Blaskey L, Rosenberg SA, Lee LC, & Levy SE (2015). Using standardized diagnostic instruments to classify children with autism in the study to explore early development. Journal of Autism and Developmental Disorders, 45(5), 1271–1280. 10.1007/s10803-014-2287-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Wiggins LD, DiGuiseppi C, Schieve L, Moody E, Soke G, Giarelli E, & Levy S (2020). Wandering among preschool children with and without autism spectrum disorder. Journal of Developmental and Behavioral Pediatrics, 41(4), 251–257. 10.1097/dbp.0000000000000780 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Yirmiya N, & Shaked M (2005). Psychiatric disorders in parents of children with autism: A meta-analysis. Journal of Child Psychology and Psychiatry and Allied Disciplines, 46(1), 69–83. 10.1111/j.1469-7610.2004.00334.x [DOI] [PubMed] [Google Scholar]
  59. Zeileis A (2006). Object-Oriented Computation of Sandwich Estimators. Journal of Statistical Software, 16(9), 1–16. 10.18637/jss.v016.i09 [DOI] [Google Scholar]
  60. Zeileis A, Köll S, & Graham N (2020). Various Versatile Variances: An Object-Oriented Implementation of Clustered Covariances in R. Journal of Statistical Software, 95(1), 1–36. 10.18637/jss.v095.i01 [DOI] [Google Scholar]

RESOURCES