Abstract
Video Abstract
OBJECTIVES: To estimate the national prevalence of parent-reported autism spectrum disorder (ASD) diagnosis among US children aged 3 to 17 years as well as their treatment and health care experiences using the 2016 National Survey of Children’s Health (NSCH).
METHODS: The 2016 NSCH is a nationally representative survey of 50 212 children focused on the health and well-being of children aged 0 to 17 years. The NSCH collected parent-reported information on whether children ever received an ASD diagnosis by a care provider, current ASD status, health care use, access and challenges, and methods of treatment. We calculated weighted prevalence estimates of ASD, compared health care experiences of children with ASD to other children, and examined factors associated with increased likelihood of medication and behavioral treatment.
RESULTS: Parents of an estimated 1.5 million US children aged 3 to 17 years (2.50%) reported that their child had ever received an ASD diagnosis and currently had the condition. Children with parent-reported ASD diagnosis were more likely to have greater health care needs and difficulties accessing health care than children with other emotional or behavioral disorders (attention-deficit/hyperactivity disorder, anxiety, behavioral or conduct problems, depression, developmental delay, Down syndrome, intellectual disability, learning disability, Tourette syndrome) and children without these conditions. Of children with current ASD, 27% were taking medication for ASD-related symptoms, whereas 64% received behavioral treatments in the last 12 months, with variations by sociodemographic characteristics and co-occurring conditions.
CONCLUSIONS: The estimated prevalence of US children with a parent-reported ASD diagnosis is now 1 in 40, with rates of ASD-specific treatment usage varying by children’s sociodemographic and co-occurring conditions.
- ADD —
- attention-deficit disorder
- ADDM —
- Autism and Developmental Disabilities Monitoring Network
- ADHD —
- attention-deficit/hyperactivity disorder
- ASD —
- autism spectrum disorder
- CI —
- confidence interval
- EBD —
- emotional, behavioral, and developmental condition
- FPL —
- federal poverty level
- NHIS —
- National Health Interview Survey
- NSCH —
- National Survey of Children’s Health
- PRR —
- prevalence rate ratio
What’s Known on This Subject:
Previous studies over the last 20 years have shown an increasing prevalence of autism spectrum disorder (ASD) among US children. Moreover, families of children with ASD have reported greater health care needs and challenges compared with children with other emotional or behavioral conditions.
What This Study Adds:
In this study, we present new nationally representative data on the prevalence of ASD, reported health care challenges, and estimates on ASD-specific behavioral and medication treatments. The estimated prevalence of US children with parent-reported diagnosis of ASD is now 1 in 40.
Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder characterized by persistent deficits in social communication and interactions and restricted, repetitive patterns of behaviors or interests.1 The prevalence of diagnosed ASD has increased in the United States and globally within the last 30 to 40 years.2–17 Although not fully understood, this increase likely results from multiple factors including broadening diagnostic criteria, increased provider ascertainment at earlier ages, increased parent awareness, and an increase in some risk factors such as births to older parents.18–20
The challenges of ASD are many and varied. Compared with the general population, children with ASD experience an increased prevalence of co-occurring conditions, such as asthma, eczema, gastrointestinal disturbances, and seizures.21 In addition, 83% of children with ASD have a co-occurrence of ≥1 non-ASD developmental diagnosis.22 Children with ASD have greater health services needs, including therapy,23 emergency department care,24 physician visits, and hospitalizations.25 Increased and unpredictable needs for health care visits can also affect parents’ employment,26 increase financial and time burdens,27 and disrupt family routines.28 Few national studies have compared the disparities in health services and challenges for families having a child with ASD.27 Although ASD currently has no known cure, the most common treatments to ameliorate the symptoms include behavioral, language, speech, physical, and occupational therapies.29 Pharmacological agents have been Food and Drug Administration–approved to treat irritability associated with ASD symptoms.30 The costs of caring for a child with ASD in the United States, including health care and non–health care services, was estimated at $17 081 per year beyond the costs of caring for a child without ASD, with total societal costs of caring for children with ASD estimated at $11.5 billion in 2011.31
The 2016 National Survey of Children’s Health (NSCH), a nationally and state-representative survey of 50 212 children, ages 0 to 17 years, offers the opportunity to address certain gaps in our knowledge. This study provides the most recent nationally representative estimate on children with ASD. Although the National Health Interview Survey (NHIS) can provide ASD prevalence estimates,9 the NSCH has a larger sample, and can explore other aspects of the child’s condition, including ASD-specific treatments, plus an in-depth examination of their health care needs and experiences. The 2016 NSCH also offers the first opportunity to provide national estimates on ASD-specific drug and behavioral treatments.
Methods
The 2016 NSCH was designed, directed, and funded by the Health Resources and Services Administration’s Maternal and Child Health Bureau. The survey was conducted by the US Census Bureau, which used a national address-based sample for data collection by either Web or mail. The NSCH provides information on the health and well-being of children on the basis of information from their parents or other caregivers (hereafter referred to as “parents”). Questionnaires were available in English or Spanish. The 2016 NSCH used a 2-phased data collection approach: (1) an initial household screener assessed the presence of children, their basic demographic characteristics, and special health care need status and (2) a substantive, age-specific topical questionnaire completed by parents of one randomly selected child per household. Children with special health care needs had a higher probability of selection compared with other children to increase the sample size among this population.
From June 2016 to February 2017, topical questionnaires were completed. The proportion of households known to include children that completed the topical questionnaire was 69.7%. The overall weighted response rate, which includes nonresponse to the screener to identify whether households include children, was 40.7%. Additional details about the survey methodology are available elsewhere.32 This study was a secondary analysis of publicly available data. The data were collected under Title 13, US Code, Section 8(b). All data products are reviewed for adherence to privacy protection and disclosure avoidance guidelines by the Census Bureau’s Disclosure Review Board.
Analyses for this study were limited to 43 283 children aged 3 to 17 years and excluded children ≤2 years (n = 6929). Parents were asked if a doctor or other health care provider had ever told them that their child had “Autism or Autism Spectrum Disorder, including diagnoses of Asperger’s Disorder or Pervasive Developmental Disorder.” Parents who responded “Yes” were subsequently asked if their child currently has the condition (see Fig 1). For this analysis, children identified as having ASD were those with a parent report of (1) ever told by a doctor and/or other health care provider that their child had ASD and (2) child currently has ASD. In 87 cases, the parent reported that the child was ever diagnosed with ASD but did not currently have the condition (7.1%). These cases were not considered as having ASD for this analysis. Children whose parents did not know or refused to answer either of the ASD questions were excluded from analyses (n = 262; 0.6%)
Flow diagram of survey-participant progress through the autism and ASD questions, NSCH, 2016.
The prevalence of ASD was estimated overall and by selected demographic, socioeconomic, and childbirth characteristics in Table 1. We also provided unadjusted and adjusted prevalence rate ratios (PRRs). In Table 2 we examined (1) health care service use including medical and behavioral treatments for ASD symptoms and receipt of specialist care other than mental health treatment and/or counseling; and (2) parental frustration, access to treatment, and quality of care received including receipt of care in a medical home. We defined medical home using the American Academy of Pediatrics framework, specifically whether33 (1) the child had a personal physician or nurse; (2) the child had a usual place for sick care; (3) the family reported no problems obtaining needed referrals; (4) the family reported receipt of family centered care; and (5) the family reported receipt of effective care coordination, if needed. All 5 criteria needed to be met for a child to have a medical home.
Point Prevalence of Parent-Reported Current ASD Among Children Aged 3–17 Years According to Selected Demographic, Socioeconomic, and Birth Characteristics, United States, 2016
PRRs of Having Selected Health Care Characteristics According to Parent-Reported ASD Status and Other EBDs, United States, 2016
We additionally explored how the health care experiences of children with ASD and their families differed from children with other diagnosed emotional, behavioral, and developmental conditions (EBDs) (attention-deficit/hyperactivity disorder [ADHD], anxiety, behavioral and/or conduct problems, depression, developmental delay, Down syndrome, intellectual disability, learning disability, Tourette syndrome), and those without ASD or other EBDs. EBDs were identified analogously to ASD: affirmative responses to both questions of whether the parent was “ever told child had condition” and whether “child currently has the condition.” Health care experiences included services used, difficulties accessing needed care, and current receipt of services under a special education or early intervention plan. We estimated weighted percentages, PRRs, and 95% confidence intervals (CIs) comparing children with ASD to children with other EBDs but not ASD, and all other children without EBDs or ASD. The PRRs were adjusted for child age, sex, and race and/or ethnicity.
Finally, we examined factors associated with the likelihood of receiving medication and/or behavioral treatment of ASD-related symptoms among children with ASD and health insurance using multivariable logistic regression. Covariates included sociodemographic characteristics, years since diagnosis (calculated by subtracting age at diagnosis from child’s current age), provider type for ASD diagnosis, and presence of co-occurring conditions. (Supplemental Table 4 provides prevalence estimates of co-occurring conditions.) Additionally, the medication model included behavioral treatment, whereas the behavioral treatment model included medication. Adjusted PRRs and 95% CIs are reported.
Missing data for child sex, race, and ethnicity were imputed by using hot deck methods during the weighting process, whereas the household income-to-poverty ratio was imputed by using regression methods. All estimates were weighted by using survey sampling weights available from the Census Bureau to produce estimates that are representative of the US noninstitutionalized child population aged 3 to 17 years. These weights reflect the inverse of the probability that the address was selected for the survey and were adjusted to account for the higher likelihood that children with special health care needs would be selected for the topical interview, as well as nonresponse. Weighted analyses were conducted by using SAS 9.4 (SAS Institute, Inc, Cary, NC) and SAS-callable SUDAAN 11.0 (RTI International, Research Triangle Park, NC).
Results
From these nationally representative data, we estimated the point-prevalence of parent-reported ASD diagnosis in 2016 to be 2.50 per 100 children. This represents an estimated 1.5 million children aged 3 to 17 years (Table 1). After adjustment for selected demographic, socioeconomic, and birth characteristics, ASD prevalence in boys was 3.46 times higher than in girls. Additionally, ASD prevalence was 47% higher for children with single mothers than children in 2-parent, married households, 2.06 times as high for children from households <100% of the federal poverty level (FPL) compared with children from households ≥400% of the FPL and 71% higher in children born preterm than term. Prevalence was 2.34 times as high for US-born than foreign-born children. The same variables were also significant when unadjusted.
Children with a parent-reported ASD diagnosis (including those who had other EBDs) had more needs and difficulties regarding health care access and use compared with children without an ASD diagnosis but who had other EBDs, and children with parent-reported ASD also had more needs and difficulties compared with other children without ASD or EBDs (Table 2). Compared with children with other EBDs, children with ASD were more likely in the past 12 months to have received mental health counseling, seen a specialist besides a mental health professional, used alternative health care or treatments, ever have a special education or early intervention plan, and currently receive special services for developmental needs. However, children with ASD were also more likely to have parents who reported difficulties with health care. They were 44% more likely to report problems getting mental health treatment, 46% more likely to report not receiving needed mental health care, 2.02 times more likely to report being usually or always frustrated in getting services, 23% less likely to have a medical home, and 24% less likely to receive needed care coordination. Children with ASD also had greater needs and difficulties compared with all other children (without other EBDs).
Of children with a parent-reported ASD diagnosis, ∼27% were taking medication for ASD-related symptoms. Children with ASD from households where the highest level of education was less than high school were 74% more likely to be currently taking medication for ASD symptoms compared with children from households with at least a college degree (Table 3). In addition, current medication use varied regionally and was higher among children who were diagnosed >7 years ago, who had attention-deficit disorder (ADD) and/or ADHD or behavioral and/or conduct problems, and who received behavioral treatment of their ASD symptoms in the past 12 months. Children in 2-parent, unmarried households were less likely than children in 2-parent, married households to be taking medication for ASD.
Percentage of Children Aged 3–17 Years Who Took Medication for ASD or Received Behavioral Treatment of ASD, Among Children Aged 3–17 Years Who Currently Have ASD and are Insured, United States, 2016
Approximately 64% of children with a parent-reported ASD diagnosis received behavioral treatment in the past 12 months. Children with behavioral and/or conduct problems had a higher prevalence of behavioral treatment, as well as those currently taking medications for ASD-related symptoms. US-born children or those diagnosed with ADD and/or ADHD were less likely to have received behavioral treatment.
Discussion
We used the recently released 2016 NSCH to estimate a nationally representative prevalence for children with a parent-reported ASD diagnosis of 2.50%. This is the fourth ASD prevalence report from the NSCH; however, because of several notable updates to the NSCH data collection, comparisons of the ASD prevalence estimates presented here with previously published NSCH estimates must be done cautiously because we cannot tell what proportion was explained by internal survey changes rather than external factors. The 2016 NSCH was conducted by using an address-based sample. Families responded by either mail or the Internet, whereas previous surveys were administered by telephone. In addition, question wording was changed twice. In the 2007 survey, the wording was expanded to add specificity and align with the Diagnostic and Statistical Manual of Mental Disorders fourth edition.1 Also, beginning with the 2007 survey, the ASD case definition was more strictly defined on the basis of parents’ affirmative responses to 2 questions (“ever diagnosed with ASD” and “currently has ASD”) to exclude diagnoses that may have been lost because of maturation, treatment, or new information.34 Questions were also modified slightly in the 2016 NSCH to reflect changes in the most recent American Psychiatric Association criteria.1
In addition to the NSCH, population-based ASD prevalence estimates for US children have been reported from the NHIS, a nationally representative in-person household survey that includes ASD questions similar to those in the NSCH,5,9,17 and the Autism and Developmental Disabilities Monitoring Network (ADDM), an ongoing surveillance system in local population-based areas in which ASD cases are identified through education and health records review.35 The most recent published NHIS ASD prevalence estimate (2.76%) is higher than that reported here but is based solely on ever receiving a diagnosis.36 The estimate of current ASD from the 2016 NHIS is 2.47%, which virtually matches the NSCH estimate reported here.37
The most recent composite ADDM prevalence estimate (1.68%), using 2014 data, was higher than the previous estimate of 1.46% using 2012 data, although this was still lower than the ASD estimate from the 2016 NSCH.35,38 Any conclusions from this latter comparison must be tempered given that estimates from these different systems reflect different years, populations (11 local US populations versus the entire United States), and ages (children aged 8 years in ADDM vs 3–17 years in NSCH).35 The estimates derived from previous versions of the NSCH and ADDM are similar when comparing overlapping years.7,39 In addition, the estimate from the NSCH falls within the range of estimates (1.31%–2.93%) from the 11 ADDM sites, and sites able to access both health and education records had higher prevalence estimates compared with sites accessing only health records.
Because there is no biological marker, ASD is a particularly challenging condition to track; thus, multiple systems with different case ascertainment strategies and supplemental data collection for children with ASD are useful in developing a full picture of ASD prevalence. Findings from the NHIS, NSCH, and ADDM each contribute unique information that, when combined, helps form a comprehensive picture of ASD among children in the United States.
Estimates reported here indicate less variation in prevalence rates across child age, race and/or ethnicity, or socioeconomic groups than observed in earlier ASD prevalence studies.7,40 Policy changes such as the 2007 AAP recommendations for universal screening by 18 to 24 months may have helped to increase ASD diagnosis among young children, thus reducing the prevalence disparity by age.41 These types of recommendations might have also had broader impacts by increasing provider and parent awareness of ASD generally, possibly contributing to increased diagnoses in traditionally underserved racial-ethnic groups. In addition, among the 87 cases of ever but not current autism, we do not know if the children achieved optimal outcomes or were initially misdiagnosed.42
The current study demonstrates that families face challenges in accessing healthcare services. Because children with ASD are likely to need multiple types of services and accompanying care coordination,43–46 the challenge of achieving a medical home appears evident among children with ASD in that they were less likely to meet the medical home criteria than other children. Although children with ASD were more likely to have a personal doctor or nurse and have a usual source for sick care, they were less likely to get needed referrals for specialty care. Consistent with previous studies, we found that children with ASD also had high rates of co-occurring mental health conditions or EBD diagnoses.21,22,35 We were not able to determine if some of the co-occurring EBD diagnoses were precursor diagnoses made as part of the diagnostic trajectory of a child receiving an ASD diagnosis versus being distinct co-occurring conditions. Nonetheless, our findings indicate that children with ASD face many developmental challenges; they particularly need referrals and care coordination.
Treatments for ASD-related symptoms have primarily been focused on behavior change and/or skill building.19 Psychotropic medications have been used to relieve some emotional and behavioral symptoms, such as challenging behaviors, irritability, anxiety, and hyperactivity.47 However, there are no pharmacologic options for treating the core deficits of ASD.48 In the current study, we found that 27% of US children with ASD took medication for ASD-related symptoms and 64% received behavioral treatment. Children with ASD who had reported diagnoses of behavioral and/or conduct problems were significantly more likely to have their ASD symptoms treated with medication and receive behavioral treatment. Children with ASD who were also reported to be diagnosed with ADD and/or ADHD were 3.30 times as likely to be treated with medication for their ASD symptoms, whereas they were only 84% as likely to receive behavioral therapy. Other studies have revealed a higher rate of psychotropic medication use for children with ASD when there was a comorbid diagnosis of ADHD.49
Our study has several strengths, including its large, nationally representative sample, which enabled examination by subgroups, and the comprehensive set of questions on ASD, co-occurring conditions, and health care experiences. However, there are also limitations. The data are cross-sectional and based on parent report of diagnoses. Undiagnosed ASD could not be assessed, so the prevalence of ASD may be underestimated, especially for younger children. Parents’ reports of diagnoses were not clinically validated; however, studies have revealed a high concordance rate (93%–98%) between parent reports of their child receiving a definite ASD diagnosis and clinician’s diagnosis in verbal children.50,51 An ASD registry in the United Kingdom revealed the reliability of parent-reported ASD diagnoses of children was 96% when compared with clinical reports.52 In addition, there was a high level of agreement between parents and clinicians on ASD-related behaviors at 12 months of age.53 However, family characteristics may affect the degree of parent-clinician agreement.54 Studies have also revealed a high test-retest reliability on maternal report of their child’s mental health conditions.55 Other studies found good agreement between parental report and pediatricians’ records,56 and fair-to-excellent agreement between parents and day care providers on recall of early symptoms associated with ASD.57 Although the specific autism diagnosis question from NHIS and NSCH has not been externally validated, the consistency of results from the NSCH, NHIS, and ADDM when compared from the same approximate period suggests a degree of reliability of these estimates.5 There is also the potential for nonresponse bias. The Census Bureau applied a nonresponse weighting adjustment that significantly mitigated any identified differential nonresponse.58 However, there could be additional sources of bias, such as selection bias, that are not controlled by the weighting adjustments. In addition, the NSCH applied poststratification adjustments to ensure that sociodemographic subgroups were appropriately represented in the estimates.
From the 2016 NSCH data, we estimated that 1 in 40 children in the United States have a parent-reported ASD diagnosis. Because ASD is a lifelong condition for most children, an important area of future research would be to study life course development and understand what factors influence health and well-being in young adulthood and beyond for these children.
Acknowledgment
We thank Alicia A. Livinski, MPH, MLS, at the National Institutes of Health Library, for her dedication and thorough library research work.
Footnotes
- Accepted August 27, 2018.
- Address correspondence to Michael D. Kogan, PhD, Office of Epidemiology and Research, Health Resources and Services Administration, Maternal and Child Health Bureau, 5600 Fishers Lane, Room 18N128, Rockville, MD 20857. E-mail: mkogan{at}hrsa.gov
The views expressed in this article are those of the authors and do not necessarily reflect the official policies of the US Department of Health and Human Services, the Health Resources and Services Administration, or the Centers for Disease Control and Prevention, nor does mention of the department or agency names imply endorsement by the US government.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Dr Shattuck is supported by grant R01MH104423 from the National Institute of Mental Health. Funded by the National Institutes of Health (NIH).
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2018-2695.
References
- Copyright © 2018 by the American Academy of Pediatrics
In this issue
Jump to section
Related Articles
Cited By...
- The social patterning of autism diagnoses reversed in California between 1992 and 2018
- Providing Inpatient Medical Care to Children With Autism Spectrum Disorder
- Increasing Early Childhood Screening in Primary Care Through a Quality Improvement Collaborative
- Individuals with autism have no detectable deficit in neural markers of prediction error when presented with auditory rhythms of varied temporal complexity
- Maternal and Child Health Bureaus Autism Research Program
- Service Use Classes Among School-aged Children From the Autism Treatment Network Registry
- The Value of State Differences in Autism When Compared to a National Prevalence Estimate