Association between autism spectrum disorder and intentional self-harm

Abstract

Background

Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by persistent challenges in communication and social interaction and, often accompanied by restricted and repetitive patterns of behavior and interests. The reported prevalence of ASD in the United States has tripled in the past two decades. Recent studies indicate that ASD is associated with increased self-injurious behaviors. The purpose of this study is to assess the excess risk of intentional self-harm associated with ASD.

Methods

Using a repeated cross-sectional study design, we analyzed data from the 2016–2020 Nationwide Emergency Department Samples (NEDS), the largest all-payer emergency department (ED) database in the United States. ED visits for intentional self-harm were identified using the ICD–10–CM external cause-of-injury matrix. Adjusted odds ratios (aORs) and 95% confidence intervals (CIs) of ED-treated intentional self-harm associated with ASD in the presence or absence of co-occurring attention-deficit hyperactivity disorder (ADHD) and/or intellectual disability (ID) were estimated through multivariable logistic regression.

Results

The 2016–2020 NEDS recorded an unweighted total of 159,590,866 ED visits, of which 2,570,446 (1.6%) were related to intentional self-harm. Using weighted data, intentional self-injury accounted for 2.3% of ED visits made by patients with a diagnosis of ASD, 3.9% of ED visits by patients with a diagnosis of ADHD, and 3.3% of ED visits by patients with a diagnosis of ID. Compared to patients without ASD or ADHD/ID, patients with ASD alone had a 65% increased odds of intentional self-harm (aOR = 1.65; 95% CI: 1.60, 1.70); in addition, patients with ADHD/ID but no ASD a 186% increased odds (aOR = 2.86; 95% CI: 2.83, 2.88), and patients with both ASD and ADHD/ ID a 170% increased odds (aOR = 2.70; 95% CI: 2.58, 2.82) of intentional self-harm. Poisoning accounted for 82.3% of the intentional self-harm-related ED visits among patients without ASD and 61.0% of intentional self-harm-related ED visits among patients with ASD.

Conclusions

ASD is associated with a significantly increased risk of ED-treated intentional self-harm, particularly in patients with co-occurring ADHD or ID. Poisoning from psychotropic and other pharmaceutical drugs is the leading mechanism of intentional self-harm.

Background

Autism spectrum disorder (ASD) is characterized by persistent differences and challenges in social communication and social interaction and commonly restricted and/or repetitive behaviors [1]. About 1 in 36 children have ASD according to estimates from Center for Disease Control’s Autism and Developmental Disabilities Monitoring (ADDM) Network [2]. Among the unique health challenges facing people with ASD is the excess risk of injury morbidity and mortality. Several epidemiological studies based on emergency department visit data have shown that children with ASD are at elevated risk of specific injuries [3–8]. People with ASD are also at heightened risk of injury mortality, with a risk of premature death 2- to 10-fold higher than in the general population. Self-harm contributes to this excess injury morbidity and mortality, as people with ASD are at increased risk of self-injurious behavior, suicidal ideation, and suicide [8–12].

Several factors contribute to the increased risk of self-harm in people with ASD, including common co-occurring mental health conditions that are associated with an increased risk of self-harm and suicide [13–15]. Prevalence estimates show that 28% of people with ASD have attention deficit hyperactivity disorder (ADHD), 20% have anxiety disorders, and 11% have depressive disorders [16]. These diagnoses are associated with increased risk of suicide and may contribute to increased prevalence of self-harm in people with ASD. Intellectual disability is also a common co-occurring diagnosis associated with self-harm and self-injurious behaviors [17]. Among children with ASD, 37.8% have intellectual disability (ID) [2]. Self-harm includes a spectrum of injuries and/or behaviors. Self-injurious behavior (intentional and unintentional), such as hand hitting, self-cutting, and hair pulling, is common in people with ASD, with an estimated prevalence of 42% [18]. There is a known link between self-injurious behavior and suicide, in people with and without ASD [10, 19–26]. In a survey of people with ASD, every 1 point increase on the suicide item of the Suicide Behaviors Questionnaire was associated with a 2.2-fold increase in the risk of self-harm [19]. While self-harm is more challenging to classify in younger children, self-injurious behaviors, such as head-banging, have been reported in children with ASD under 3 years of age [27]. We examined intentional self-harm incidents presenting to US EDs using the Nationwide Emergency Department Sample (NEDS) to explore epidemiologic factors and common co-occurring diagnoses.

Methods

Data for this study came from the 2016 to 2020 NEDS, the largest all-payer emergency department database in the US [28]. The NEDS is a nationally representative sample of over 28 million ED visits each year which translates to 123 million annual hospital-owned ED visits in the US. The NEDS is developed for the Healthcare Cost and Utilization Project through a federal-state-industry partnership sponsored by the Agency for Healthcare Research and Quality.

Data elements compiled in the NEDS come from state inpatient databases (patients initially seen in the ED and admitted to the same hospital) and state emergency department databases (patients released from ED or transferred to other hospitals). Using the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM), the NEDS presents up to 35 diagnoses associated with each visit in addition to visit type and demographic, hospital, and regional characteristics. Race and ethnicity data are not available prior to 2019. In a sensitivity analysis, 2019–2020 data controlling for race and ethnicity were compared to findings from the overall 2016–2020 dataset.

Study design

A repeated cross-sectional study design was used to assess the excess risk of self-injury associated with ASD. Analyses included patients across the lifespan, excluding infants less than 1 year old, as ASD diagnosis is less reliable in this age group.

Measures

Intentional self-harm was identified using ICD10 codes X60–X84, ASD using ICD10 code F84.0, ADHD using ICD10 code F90.9, and intellectual disability (ID) using ICD10 codes F70-F73. Mechanism of injury was coded according to CDC external cause of injury codes. Age was categorized into 11 groups: 1–4, 5–9, 10–14,15–19, 20–24, 25–29, 30–34, 35–44, 45–54, 55–64, and ≥ 65 years. Demographic variables analyzed include age, sex, and regional location of emergency department.

Statistical analyses

The prevalence of ED visits for self-harm was estimated and stratified according to age, sex, region, ASD, ADHD, and ID. We used the Rao-Scott approach to obtain weighted frequencies [29]. Unadjusted and adjusted odds ratios (ORs) and 95% confidence intervals (CIs) from logistic regression models were used to estimate the association of demographic characteristics and comorbidities with self-harm. In addition, ORs and 95% CIs of intentional self-harm associated with ASD were estimated according to the presence of ADHD and/or ID. Weighted frequencies of mechanism of injury and type of poisoning were computed according to ASD status.

Results

From January 1, 2016 to December 31, 2020, the NEDS recorded a total of 159,590,866 ED visits; of these, 2,570,446 (1.6%) were related to intentional self-harm. This yielded a weighted frequency of ED visits in the US of 684,346,447 during the five-year study period. An estimated 10,941,710 ED visits (2.5%) were related to intentional self-harm, representing an annualized rate of 6.6 ED visits per 1000 population. The prevalence of intentional self-harm in ED visits varied significantly with demographic and clinical characteristics (Table 1). Intentional self-harm accounted for 0.7% of ED visits made by children aged 5–9 years, 1.3–1.9% of ED visits by adolescents and adults aged 10–64 years, and 2.2% of ED visits by older adults (Table 1). Overall, the prevalence of intentional self-harm in ED visits was slightly higher among male patients than among female patients (1.7% vs. 1.6%; p < 0.001), and was highest in the Midwest (1.8%), followed by the West (1.7%), the Northeast (1.6%) and the South (1.5%) (Table 1). Intentional self-harm accounted for 2.3% of ED visits made by patients with a diagnosis of ASD, 3.9% by patients with a diagnosis of ADHD, and 3.3% by patients with a diagnosis of ID (Table 1), with an overall national prevalence of 1.6% (Table 1). Most patients were transferred or admitted (54.1%), 44.6% were discharged, and 0.1% died (Table 2). A larger proportion of patients with ASD were discharged from the ED, in comparison to patients without ASD (Fig. 1). In a sensitivity analysis to determine the effects of race and ethnicity, the odds of ED visit for intentional self-harm among people with ASD, adjusted for age, sex, ASD, ADHD, ID, and region from 2016 to 2020 in comparison to people without ASD (aOR 1.37, 95% CI: 1.34, 1.40), are similar to the restricted analysis from 2019 to 2020 adjusted for race and ethnicity (in addition to age, sex, ASD, ADHD, ID and region; aOR 1.30, 95% CI: 1.12, 1.45).

Table 1.

Weighted frequencies, estimated odds ratios (ORs) and 95% confidence intervals (CIs) of intentional self-harm according to demographic and clinical characteristics, weighted data from the Nationwide Emergency Department Sample 2016–2020

Characteristic	Number of ED visits (n = 684,346,447)	Number of ED visits for self-harm (n = 10,941,710) (%)	Crude OR (95% CI)	Adjusted* OR (95% CI)
Age (years)
1–4	39017100	396231 (1.02)	1.55 (1.54, 1.57)	1.60 (1.58, 1.62)
5–9	28394700	186697 (0.66)	1.00	1.00
10–14	25517394	323038 (1.27)	1.94 (1.91, 1.96)	1.88 (1.86, 1.90)
15–19	38050311	691795 (1.82)	2.80 (2.77, 2.83)	2.79 (2.76, 2.82)
20–24	50711421	692470 (1.37)	2.09 (2.07, 2.11)	2.14 (2.11, 2.16)
25–29	54456238	742014 (1.36)	2.09 (2.06, 2.11)	2.14 (2.12, 2.17)
30–34	50592685	681800 (1.35)	2.06 (2.04, 2.09)	2.12 (2.10, 2.15)
35–44	86854767	1201450 (1.38)	2.12 (2.10, 2.14)	2.18 (2.16, 2.20)
45–54	84343677	1352904 (1.60)	2.46 (2.44, 2.49)	2.54 (2.51, 2.56)
55–64	83120954	1567138 (1.89)	2.90 (2.87, 2.93)	2.99 (2.96, 3.02)
≥65	143287200	3106174 (2.17)	3.35 (3.31, 3.38)	3.46 (3.43, 3.50)
Sex
Female	378063623	5853315 (1.55)	1.00	1.00
Male	306228388	5086451 (1.66)	1.07 (1.07, 1.08)	1.08 (1.07, 1.08)
Region
Northeast	124086524	1979163 (1.60)	1.08 (1.08, 1.09)	1.06 (1.05, 1.06)
Midwest	154849381	2707737 (1.75)	1.19 (1.18, 1.19)	1.17 (1.17, 1.18)
South	276242142	4079652 (1.48)	1.00	1.00
West	129168400	2175157 (1.68)	1.14 (1.14, 1.15)	1.13 (1.13, 1.13)
Autism spectrum disorder
No	682989595	10910285 (1.60)	1.00	1.00
Yes	1356851	31425 (2.32)	1.47 (1.43, 1.50)	1.35 (1.32, 1.39)
Attention-deficit/hyperactivity disorder
No	680146853	10776486 (1.58)	1.00	1.00
Yes	4199593	165224 (3.93)	2.54 (2.52, 2.57)	2.86 (2.83, 2.89)
Intellectual disability
No	683328383	10896167 (1.60)	1.00	1.00
Yes	1018064	45543 (3.33)	2.12 (2.08, 2.17)	2.35 (2.30, 2.40)

*Adjusted for age, sex, ASD, ADHD, ID, and region

Table 2.

Disposition after ED visit for self-harm and non-self-harm visits

Disposition	Type of ED Visits		Total No. (%)
	Self-harm No. (%)	Non self-harm No. (%)
Discharged	4,874,138 (44.55)	546,150,044 (81.10)	551,024,182 (80.52)
Transferred	745,765 (6.82)	22,456,961 (3.33)	23,202,726 (3.39)
Left against medical advice	139,442 (1.27)	10,345,643 (1.53)	10,485,085 (1.53)
Admitted	5,167,490 (47.23)	93,461,778 (13.88)	98,629,268 (1.41)
Died	14,874 (0.14)	990,312 (0.15)	1,005,186 (0.15)
Total	10,941,710 (100.00)	673,404,737 (100.00)	684,346,447 (100.00)

Fig. 1.

Disposition of intentional self-harm ED visits among people with and without ASD. Blue square represents ED visits for intentional self-harm among patients with ASD. Orange square represents ED visits for intentional self-harm among patients without ASD

The distribution of self-harm in ED visits by age differed between patients with ASD and those without ASD (Fig. 2). Among patients with ASD, the prevalence of intentional self-harm increased with age during childhood and adolescence and remained at about 3% until age ≥ 65 years; among patients without ASD, the prevalence showed a multimodal distribution with peaks at ages 15–19 and ≥ 65 years (Fig. 2).

Fig. 2.

Prevalence of intentional self-harm in emergency department visits according to age and ASD status. White circles represent prevalence of self-harm ED visits among people with ASD. Black boxes represent prevalence of self-harm ED visits among people without ASD

Poisoning was the leading mechanism of intentional self-harm in both patients with ASD and those without ASD. Of the 31,425 intentional self-harm-related ED visits made by patients with ASD, 61.0% were due to poisoning, 13.2% due to cutting or piercing, 8.1% due to firearms, 6.1% to traffic injury (e.g., jumping or lying in front of a car), and 11.6% to other mechanisms (e.g., fire or burn, jumping from height, natural/environment, and suffocating). Among patients without ASD, the most common mechanisms for ED-treated incidents of intentional self-harms were poisoning (82.3%) and natural/environment (7.9%) and least commonly by cutting/piercing (4.2%), and traffic injury (0.5%) (Fig. 3).

Fig. 3.

Percentage distributions of injury mechanisms by ASD status in intentional self-harm related ED visits. Injury mechanisms by autism spectrum disorder (ASD) status in intentional self-harm related ED visits, weighted data from the Nationwide Emergency Department Sample 2016–2020

Among patients with ASD, psychotropic drugs were the substances most commonly involved in ED visits for self-harm due to poisoning, accounting for 20.8% of the poisoning cases, followed by other systemic drugs (16.5%), and antiepileptic, sedative-hypnotics (16.2%). The remaining poisoning cases involved a variety of substances, including medications such as systemic antibiotics. Among patients without ASD, systemic antibiotics, anti-infectives, and hematological agents were involved in 19.2% of ED visits for self-harm due to poisoning, followed by other substances (15.1%), and illicit drugs (14.8%), (Table 3).

Table 3.

Weighted Nationwide Emergency Department Sample 2016–2020, percentage distributions of poisoning types by autism spectrum disorder (ASD) status in intentional self-harm discharges

Type of poisoning	ASD n (%)	No ASD n (%)
Systemic antibiotics, anti-infectives, and hematological agents	2,579 (12.1)	1,923,944 (19.2)
Psychotropic drugs	4,433 (20.8)	631,294 (6.3)
Illicit drugs (narcotics and hallucinogens)	597 (2.8)	1,483,040 (14.8)
Other pharmaceutical drugs
Antiepileptic, sedative- hypnotics	3,453 (16.2)	711,458 (7.1)
Hormones and their synthetic substitutes	980 (4.6)	1,012,075 (10.1)
Diuretics	2,664 (12.5)	1,332,732 (13.3)
Other systemic drugs	3,517 (16.5)	1,412,896 (14.1)
Other substances	3,090 (14.5)	1,513,102 (15.1)
Total	21,313 (100.0)	10,020,541 (100.0)

Compared to patients without ASD, ADHD, or ID, those with ASD alone had a 65% increased odds of intentional self-harm, patients with ADHD and/or ID but without ASD had a 186% increased odds, and patients with both ASD and ADHD and/or ID had a 170% increased odds of intentional self-harm (Table 4). The interaction between ASD and ADHD and/or ID on the odds of intentional self-harm was negative and statistically significant on both the multiplicative (b=−0.56, p < 0.0001) and additive scales (RERI = −0.81, 95% CI: −0.94, −0.67 AP = −0.30, 95% CI: −0.36, −0.24; S = 0.68, 95% CI: 0.63, 0.73).

Table 4.

Nationwide Emergency Department Sample 2016–2020 weighted odds ratios (ORs) and 95% confidence intervals (CIs) of self-harm associated with autism spectrum disorder (ASD) according to attention deficit hyperactivity disorder (ADHD) and/or intellectual disability (ID) status

Comorbidity status		Percentage of ED visits for self-harm	Crude OR (95% CI)	Adjusted* OR (95%CI)
Autism spectrum disorder	Attention-deficit/hyperactivity disorder and/or Intellectual disability
No	No	2.48	1.00	1.00
Yes	No	3.28	1.28 (1.25, 1.32)	1.65 (1.60, 1.70)
No	Yes	5.80	2.59 (2.56, 2.61)	2.86 (2.83, 2.88)
Yes	Yes	5.03	2.27 (2.17, 2.37)	2.70 (2.58, 2.82)

*Adjusted for age, sex, and region

Discussion

This national study assessed the epidemiology of ED presentations for intentional self-harm among people with ASD across the lifespan. Results from this analysis indicate that people with ASD more commonly visit the ED for intentional self-harm when compared to people without a diagnosis of ASD. This is consistent with prior literature showing increased poisonings [30] and suicidal thoughts and behaviors [8] among people with ASD. In a meta-analysis encompassing people across the age spectrum, ASD was associated with an increased pooled odds (3.18; 95% CI, 2.45–4.12) of self-injurious behavior and suicidality in children and adults [8]. While our study also reported increased odds of intentional self-harm in ASD patients, those with ASD alone had the lowest odds at 1.65, while patients with ASD and concomitant ADHD and/or ID had odds more consistent with that reported in the meta-analysis. Our study included self-harm warranting an ED visit and thus doesn’t capture thoughts of self-harm or self-injurious behaviors or injuries that do not warrant acute medical care. Prior NEDS analyses from 2008 investigating injury-related ED visits among children 3–17 years, found higher odds of self-harm among children with autism (OR: 5.40, 95% CI: 4.18–6.92, p < 0.05) [6]. As this study only included children and our study includes people with ASD across the lifespan, the difference may, in part, be due to the known linear positive cohort effect in autism diagnoses [31]. Adult patients and specifically older adults are less likely (due to evolved diagnostic criteria, diagnostic awareness, and other factors) to be diagnosed with autism. This phenomenon may also contribute to our finding of a larger difference in the proportion of ED visits for intentional self-harm among young people with ASD in comparison to children and adolescents from the general population than among adults > = 65. The proportion of ED visits for intentional self-harm in children and adults without ASD followed an expected multi-modal distribution. In comparison, children and adults with ASD had an elevated proportion of visits from ages 5 to 64 years. At the extremes of age there is likely greater misclassification in autism diagnoses as the significant differences in proportion of visits otherwise persists across the age spectrum.

Mechanisms of intentional self-harm differed between people with and without ASD. Poisoning was the most common mechanism of ED-treated intentional self-harm in both those with and without ASD, accounting for a larger proportion of visits among those without ASD. Children and adults with ASD were more likely to present to the ED after poisoning with psychotropic drugs and anti-epileptics and less likely to present to the ED after poisoning from illicit drugs. This is consistent with increased access to psychotropic and antiepileptic medication among people with autism [32] and prior literature that explored all poisoning related ED visits in children with autism [30]. People with ASD are at a lower risk of substance use disorder, which may contribute to the lower proportion of intentional self-harm visits for poisoning [33, 34].

In comparison to people without ASD, ED-treated incidents of intentional self-harm in people with ASD were more frequently caused by cutting or piercing, firearms, and motor vehicle-related incidents (Fig. 3). Higher proportions of these three mechanisms of self-harm are accounted for, in part, by the relatively lower proportion harming themselves by poisoning. The higher proportion of intentional self-harm via cutting or piercing likely reflects the association between ASD and non-suicidal self-harm and self-injury (both intentional and unintentional) [8, 35, 36]. Children with ASD are potentially more resistant to child passenger safety restraints and may be at higher risk for throwing themselves from vehicles [37]. Children with ASD are less likely to present to EDs for firearm injuries (regardless of intent) when compared to such visits in children without ASD or ID [6]. In contrast we found an increased proportion of intentional self-harm incidents evaluated in EDs associated with firearms in people with ASD. These findings reflect important differences in mechanism of self-harm and warrant further investigation to determine age-related patterns and areas for targeted prevention.

People with ASD, ADHD and/or ID were all more likely to visit the ED for intentional self-harm (Table 4). The common co-occurring psychiatric diagnoses of ADHD and ID can explain only part of the excess risk of self-injury in people with ASD. Compared to people without a diagnosis of ASD, ADHD, or ID, the estimated odds of being treated in the ED for intentional self-harm among people with a diagnosis of ASD without ADHD and/or ID were lower than the odds for people without ASD but with ADHD and/or ID. The negative interaction between ASD and ADHD and/or ID on the risk of ED visits for intentional self-harm suggests that the co-occurring disorders reduce, rather than compound, the risk of intentional self-harm typically associated with ADHD or ID alone. This significant negative interaction indicates that while people with ADHD or ID may have a higher intentional self-harm risk profile, the presence of ASD in combination with these disorders is associated with a lower-than-expected risk of ED-treated intentional self-harm. This could be due to increased supervision and structure among children and adults with complex comorbidities and/or restricted autonomy due to cognitive or functional challenges.

Limitations

This study has several limitations. The NEDS only captures intentional self-harm that warrants medical attention. People with ASD may exhibit self-harming behaviors such as head banging and biting that are not included in this data analysis. Intent may also be challenging to capture, particularly among patients who have limited verbal expression. Similarly, when patients present to acute care settings, documentation of ASD diagnosis may not occur, resulting in possible misclassification and possible underestimation of cases. Race and ethnicity data was not collected for the NEDS database prior to 2019. Due to this limitation, we were not able to explore potential differences based on race and ethnicity among people with ASD [38]. Finally, we did not explore variations in mechanisms by age among the population with ASD, which could create more targeted prevention initiatives for specific means.

Conclusions

This nationally representative study of emergency department visits from 2016 to 2020 reports the increased prevalence of ED-treated intentional self-harm among children and adults with ASD and describes epidemiologic and demographic patterns. Among children and adults ages 5–64 years, a significantly higher proportion of ED visits among people with ASD were for intentional self-harm, in comparison to people without ASD. We found that ADHD and ID were independently associated with ED visits related to intentional self-harm. Intentional self-injury is multifactorial and includes a spectrum of injuries from recurring self-injurious behaviors that may not result in significant injury to suicidal thoughts and behaviors. Considering this spectrum of presentations, behaviors, and co-occurring diagnoses in the population with ASD, future studies should explore specific types of self-harm and potential preventive interventions targeted to people with ASD in various life-stages. Among children and teens with ASD, providing families with safety planning tools, such as lethal means counseling, may be advantageous. Limited research explores morbidity and mortality among adults with ASD and the unique risk of intentional self-harm among this adult population, warranting deeper investigation of circumstances, gaps in care, and contributing factors.

Abbreviations

aOR

Adjusted odds ratio

ASD

Autism spectrum disorder

CDC

Centers for Disease Control and Prevention

CI

Confidence interval

ED

Emergency department

ICD-10-CM

International Classification of Diseases, Tenth Revision, Clinical Modification

ID

Intellectual disability

NEDS

Nationwide Emergency Department Sample

US

United States

Author contributions

AB, SC, GL, CD conceptualized the research and analysis plan. SC conducted the data analysis. AB and GL drafted the initial manuscript. CI, SC, CD revised the manuscript and provided critical insights and approved the final manuscript.

Funding

This research was supported by Grants R21 HD098522 and R01 HD108138 from the National Institute of Child Health and Human Development, National Institutes of Health. The contents of the manuscript are solely the responsibility of the authors and do not necessarily reflect the official views of the funding agency.

Data availability

Data is accessible via the Healthcare Cost and Utilization Project.

Declarations

Ethics approval and consent to participate

This study does not meet the criteria of human subjects research per 45 CFR 46.

Consent for publication

Not Applicable.

Competing interests

CD and AB are editorial board members and GL is Editor-in-Chief of Injury Epidemiology. They were not involved in the peer-review or handling of the manuscript. The authors have no other competing interests to disclose.