[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
April 2007

Health Care Expenditures Associated With Autism Spectrum Disorders

Author Affiliations

Copyright 2007 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2007

Arch Pediatr Adolesc Med. 2007;161(4):350-355. doi:10.1001/archpedi.161.4.350

Objective  To examine the health care expenditures associated with autism spectrum disorders (ASDs) in medical care settings.

Design  Retrospective analysis of health insurance claims data.

Setting  Administrative claims database for a national sample of privately insured individuals.

Participants  Children and adolescents 17 years and younger with a diagnosis of a mental disorder between 2000 and 2004.

Main Outcome Measures  Annual inpatient, outpatient, and prescription drug expenditures.

Results  Average health care expenditures for individuals with an ASD increased 20.4% from $4965 per patient in 2000 to $5979 per patient in 2004, even after adjustment for inflation. When combined with rising ASD prevalence rates, total expenditures per 10 000 covered lives associated with ASDs increased 142.1% over the 5-year period. Although total expenditures per treated patient were higher for patients with ASDs than for individuals with other mental disorders, ASDs created a smaller burden on health insurers because of their relatively low treated prevalence.

Conclusions  In light of anticipated patterns of earlier identification and more proactive treatment of ASDs in the years to come, the burden of autism on the health care system will continue to increase. Efforts should be made to ensure that access to care for this vulnerable population is not compromised.

Reports of increased prevalence of autism and other autism spectrum disorders (ASDs) have fueled concerns of an apparent autism “epidemic.”1-4 It is unclear to what extent these increases represent true increases in prevalence or instead reflect greater public awareness, better and earlier clinical identification, a perverse incentive whereby children diagnosed with an ASD are eligible for better educational services, or a combination thereof. Regardless of the underlying explanation, most studies converge on an increase in the prevalence of ASDs over the past decades.1-3,5

While a number of studies have examined the prevalence of ASDs, very little is known about the costs associated with these disorders.6 The system of care for individuals with ASDs is fragmented, with services delivered in special education settings,7 residential treatment settings,8 and traditional health care settings.9 However, very few studies have examined costs associated with treatment in these settings. The studies that have been done suffer from a number of limitations, such as small sample size,6,10 the use of data sources that are quite dated,9 and limited generalizability due to the use of samples that were not geographically representative.11,12 In addition, many previous studies of costs associated with autism are from the United Kingdom,6,8,10 and it is difficult to generalize these results to the United States because of differences in practice style and health care financing.

Understanding the costs associated with treating patients with ASDs is important for several reasons. First, there is an opportunity cost associated with treatment of ASDs; resources used to treat patients with ASDs are then not available for treating other patients. Quantifying these costs may lead to a more efficient allocation of health care resources because policy makers would have more information with which to make health care financing decisions. Second, appropriate financing of services for people with ASDs is an important component of access to care.13 If public and private health insurance programs are not designed so that coverage for people with ASDs is appropriate, access to care for these patients will be compromised. A thorough understanding of the costs associated with treating these patients is necessary to design appropriate insurance programs and thereby reduce the potentially catastrophic financial burden on families of individuals with ASDs. Finally, determining the costs associated with a condition is a necessary and indispensable first step in the process of developing policy priorities. Hence, information on treatment costs associated with autism is a highly demanded policy tool for such organizations as the Interagency Autism Coordinating Committee, which was established under the Children's Health Act of 2000,14 in its mission to coordinate research and other efforts within the Department of Health and Human Services.

The objective of this study was to determine the costs of services delivered in traditional health care settings for individuals with autism and other ASDs. Using a national health insurance claims database, we tracked inpatient, outpatient, and pharmacy expenditures for privately insured children and adolescents with an ASD diagnosis between 2000 and 2004. Expenditures were then compared with those associated with other mental disorders in order to place autism costs within the broader spectrum of mental health care.

Data source

Data for this study come from the Thomson/Medstat MarketScan database, which compiles claims information from private health insurance plans of large employers. The data set contains claims information for individuals across the United States who are insured through the benefit plans of large employers. The covered individuals include employees, their dependents, and early retirees of companies who participate in the database. Thomson/Medstat collects the claims data, standardizes and combines them, and then reports back to the firms who participate. Information about the firms is unavailable for reasons of confidentiality. The database contains information for more than 4.5 million covered lives on average between 2000 and 2004. These claims data are collected from more than 200 different insurance companies, including Blue Cross and Blue Shield plans and third party administrators. Our sample consists of all individuals in the database 17 years and younger who had a mental disorder (psychiatric or substance abuse) between 2000 and 2005 (N = 256 646). We also have supplementary information on the entire enrolled population (eg, age, sex, geographic region). The study was approved by the institutional review board of the Yale School of Medicine.


From the claims data, we first calculated measures of inpatient, outpatient, and pharmacy expenditures for each individual in the study sample in each year. Expenditures were then summed across these 3 components to derive total health care expenditures. Expenditures associated with a claim represent the actual paid amount and include patient payments (deductibles or copayments), payments made by the patient's insurance plan, and any payment by other insurance providers (subrogation and Medicare savings). All expenditures were adjusted for inflation using the medical care component of the Consumer Price Index, with all amounts reported in 2004 dollars.

Inpatient and outpatient expenditures were assigned to 1 of 11 mutually exclusive diagnostic groups on the basis of primary and secondary International Classification of Diseases, Ninth Revision (ICD-9) diagnosis codes that were associated with each claim. These groups included (ICD-9 codes in parentheses): autism (299.0), other ASDs (299.1-299.9), adjustment disorder (309.0, 309.2), anxiety disorder (300.0-300.3, 307.20-307.23, 308, 313.0-313.2), bipolar disorder (296.0-296.1, 296.4-296.81, 296.89-296.9), depression (296.2-296.3, 296.82, 300.4-300.5, 301.10, 309.1, 311), hyperactivity (314), mental retardation (317-319), psychosis (295, 297-298), substance abuse (291-292, 303-305), and other mental disorders (290-319, not elsewhere specified). With the exception of the ASDs and mental retardation, these categories have been used in previous studies of health care service use and costs for children and adolescents with mental disorders.15,16

Each patient was assigned a primary diagnosis in each year in which they received services. Patients with any claims for autism were classified in the autism group. Then, any patients with any claims for ASDs other than autism were classified in the “other ASD” group. Individuals were only assigned a diagnosis of autism or “other ASD,” but not both. In addition, patients with either an autism or “other ASD” diagnosis were also included in an “any ASD” group. The remaining patients were classified according to the diagnosis responsible for the highest proportion of health care expenditures during the year.


First, the treated prevalence of ASDs in the claims database was determined for each year and compared with the treated prevalence of other mental health conditions. Next, inpatient, outpatient, and pharmacy expenditures per treated patient were determined for patients with ASDs and then summed to compute total expenditures per treated patient. Expenditures for individuals with ASDs were compared with those for other mental disorders as well. Finally, expenditures per treated patient were multiplied by the treated prevalence estimates to generate total expenditures per 10 000 covered lives associated with ASDs. This measure of overall burden to the health care system was also compared with the burden associated with other mental disorders.


The database contained information on approximately 1 million covered lives 17 years and younger per year. Of these individuals, we identified 256 646 unique patients with a mental disorder between 2000 and 2004. Characteristics of these patients are presented in Table 1. Approximately 60% of children and adolescents with a mental disorder were male, and the average age was about 12 years. The most common mental disorders were hyperactivity and adjustment disorder. The proportion of children and adolescents with a mental disorder who were diagnosed with an ASD increased from 1.8% in 2000 to 2.9% in 2004.

Table 1. 
Characteristics of the Sample
Characteristics of the Sample

Characteristics of individuals with ASDs are presented in Table 2. Patients with an ASD were slightly younger, averaging about 9 years of age, and were predominantly male (approximately 80%). Comorbid mental disorders, defined as having any claims with an associated ICD-9 code within the relevant ranges listed earlier, were relatively uncommon.

Table 2. 
Characteristics of Patients With an ASD
Characteristics of Patients With an ASD

Table 3 reports the treated prevalence of ASDs and other mental disorders in the claims database over the study period. The treated prevalence of any ASD ranged from 9.5 per 10 000 individuals in 2000 to 19.2 per 10 000 in 2004. The prevalence of ASDs was low relative to other mental disorders; only psychosis and mental retardation had smaller prevalence, averaging 2.5 and 1.8 cases per 10 000, respectively. However, the prevalence of ASDs increased dramatically during the 5 years of the study (by 101.1%). This rate of increase was much higher than that seen for all of the other mental health diagnostic groups except bipolar disorder, which increased almost 106%.

Table 3. 
Treated Prevalence per 10 000 Individuals of Mental Health Conditions by Year in Patients 17 Years and Younger
Treated Prevalence per 10 000 Individuals of Mental Health Conditions by Year in Patients 17 Years and Younger

Expenditures, however, were relatively high among patients with ASDs, particularly for those with autism (Table 4). In 2000, total health care expenditures for patients with autism averaged $5316 per patient; only patients with bipolar disorder ($5766), psychosis ($6787), or mental retardation ($11 662) had higher average expenditures. Over time, average autism expenditures increased by 26.1% to $6706 per patient in 2004, surpassing the expenditures associated with bipolar disorder and psychosis. Only mental retardation was associated with higher expenditures per patient ($10 036), which was notably 13.9% lower than comparable expenditures in 2000. Expenditures for claims specifically associated with an autism diagnosis increased 60.6% from $2061 to $3309 per treated patient (data not shown). Similar patterns were seen with respect to median expenditures.

Table 4. 
Total Health Care Expenditures per Patient for Mental Health Disorders Among Patients 17 Years and Younger*
Total Health Care Expenditures per Patient for Mental Health Disorders Among Patients 17 Years and Younger*

Finally, Table 5 reports total health care expenditures per 10 000 covered lives. These cost estimates combine the information from the treated prevalence figures and the expenditures per treated individual to generate a measure of total burden on the health care system associated with these mental disorders. Expenditures associated with ASDs increased substantially over the study period, from $47 378 per 10 000 in 2000 to $114 710 per 10 000 in 2004, an increase of 142.1%. Expenditures associated with autism increased slightly more over time than expenditures associated with other ASDs (151.7% and 121.6%, respectively). Despite the high relative rise over time, these expenditures were small in magnitude compared with other, more prevalent mental health conditions. Hyperactivity had the highest burden, averaging $406 238 per 10 000 covered lives per year over the same period.

Table 5. 
Total Health Care Expenditures per 10 000 Covered Lives Associated With Mental Health Conditions Among Patients 17 Years and Younger*
Total Health Care Expenditures per 10 000 Covered Lives Associated With Mental Health Conditions Among Patients 17 Years and Younger*

This study is the first, to our knowledge, to provide comprehensive health care cost estimates for a national sample of children and adolescents with autism and other ASDs. Our sample included 9506 individuals 17 years and younger with a diagnosis of an ASD between 2000 and 2004, 5469 (57.6%) of whom had a diagnosis of typical autism vs other ASDs. We found that the expenditures associated with autism were high relative to other mental health conditions and increased steadily over time. In addition, consistent with other studies,1-4 we found that the reported prevalence of autism is rising, although it is still small relative to other mental disorders in children and adolescents. As a result, despite high and rising expenditures per treated patient, the burden of autism on health insurance providers is still small compared with other mental disorders.

Insurance companies often designate autism as a diagnostic exclusion, meaning that any services rendered explicitly for the treatment of autism are not covered by the plan, even if those services would be covered if used to treat a different condition.13 Despite the fact that many states have passed legislation requiring that autism be covered by health insurers,13 companies that self-insure their employees (such as those that participate in the MarketScan database) are exempt from state insurance legislation under the Employee Retirement Income Security Act (ERISA). However, it is not clear to what extent health plans in our sample excluded ASDs. Since insurance companies often designate autism as a diagnostic exclusion, we would expect to find no claims with an ICD-9 diagnosis of autism in such plans. The fact that we do find claims for autism in our sample suggests that at least some of the plans in the MarketScan database do not exclude autism entirely. However, these claims data may only provide information on a subset of children and adolescents with ASDs, perhaps those with nonspecific autism-related needs that are not easily coded as a serious emotional disturbance. Hence, to the extent that health plans in the database exclude only certain services for autism or have restrictive benefit limits on services for autism, total health care expenditures associated with autism may be underestimated.

There are a number of other limitations of the study that deserve comment. First, the generalizability of the study may be limited. The MarketScan database contains health care claims for privately insured individuals. However, many children with ASDs receive health care coverage through Medicaid.17 Although the prevalence of ASDs in the MarketScan database follows similar patterns as those seen in other populations, it is not clear whether Medicaid expenditures for treating ASDs would be similar to those found in our privately insured sample. A second limitation of the study is that we only have information on expenditures for services delivered in traditional health care settings. As noted earlier, services for children and adolescents with autism are delivered in many other settings, including the education system and residential treatment. Indeed, the costlier services for ASDs may be provided in such settings, perhaps more so than in traditional medical ones. We also have no information on family out-of-pocket costs (other than copays or deductibles), lost productivity, and caregiver burden. Previous studies suggest that these costs are likely to be substantial.6,10 Finally, administrative claims data are very limited in the clinical information that they contain. In particular, we have little information about the quality of care, severity of illness, or outcomes of treatment.

Despite these limitations, it is clear that health care expenditures for children and adolescents with autism are considerable and are increasing over time. Efforts should be taken to identify the full range of autism-related services and describe where responsibility for provision and financing currently falls in order to identify gaps that can lead to family burden and constitute barriers to care. This information is necessary for designing public policies that can fill the gaps between the fragmented system of care for individuals with ASDs and ultimately reduce financial risk for families and ensure their access to timely services.

Back to top
Article Information

Correspondence: Douglas L. Leslie, PhD, Departments of Psychiatry and Epidemiology/Public Health, Yale University School of Medicine, NEPEC/182, 950 Campbell Ave, West Haven, CT 06516 (douglas.leslie@yale.edu).

Accepted for Publication: November 15, 2006.

Author Contributions:Study concept and design: Leslie and Martin. Acquisition of data: Leslie. Analysis and interpretation of data: Leslie and Martin. Drafting of the manuscript: Leslie and Martin. Critical revision of the manuscript for important intellectual content: Leslie and Martin. Statistical analysis: Leslie and Martin. Obtained funding: Leslie. Administrative, technical, and material support: Leslie. Study supervision: Leslie and Martin.

Financial Disclosure: None reported.

Funding/Support: This study was supported by grant R01 MH073884 from the National Institute of Mental Health (primary investigator, Dr Leslie).

Fombonne  E The prevalence of autism.  JAMA 2003;28987- 89PubMedGoogle ScholarCrossref
Fombonne  E Epidemiological surveys of autism and other pervasive developmental disorders: an update.  J Autism Dev Disord 2003;33365- 382PubMedGoogle ScholarCrossref
Fombonne  E Is there an epidemic of autism?  Pediatrics 2001;107411- 412PubMedGoogle ScholarCrossref
Wing  LPotter  D The epidemiology of autistic spectrum disorders: is the prevalence rising?  Ment Retard Dev Disabil Res Rev 2002;8151- 161PubMedGoogle ScholarCrossref
Fombonne  E Epidemiology of autistic disorder and other pervasive developmental disorders.  J Clin Psychiatry 2005;66 ((suppl 10)) 3- 8PubMedGoogle Scholar
Jarbrink  KFombonne  EKnapp  M Measuring the parental, service and cost impacts of children with autistic spectrum disorder: a pilot study.  J Autism Dev Disord 2003;33395- 402PubMedGoogle ScholarCrossref
Yell  MKatsiyannis  A Critical issues and trends in the education of students with autism spectrum disorders: introduction to the special issue.  Focus Autism Other Dev Disabl 2003;18138- 139http://www.ingentaconnect.com/content/proedcw/focus/2003/00000018/00000003/art00001Google ScholarCrossref
Jarbrink  KKnapp  M The economic impact of autism in Britain.  Autism 2001;57- 22PubMedGoogle ScholarCrossref
Birenbaum  AGuyot  DCohen  HJ Health care financing for severe developmental disabilities.  Monogr Am Assoc Ment Retard 1990; ((14)) 1- 150PubMedGoogle Scholar
Curran  ALSharples  PMWhite  CKnapp  M Time costs of caring for children with severe disabilities compared with caring for children without disabilities.  Dev Med Child Neurol 2001;43529- 533PubMedGoogle ScholarCrossref
Jacobson  JWMulick  JAGreen  G Cost-benefit estimates for early intensive behavioral intervention for young children with autism—general model and single state case.  Behav Interven 1998;13201- 226.http://www3.interscience.wiley.com/cgi-bin/jissue/10050475Google ScholarCrossref
Mandell  DSCao  JIttenbach  RPinto-Martin  J Medicaid expenditures for children with autistic spectrum disorders: 1994 to 1999.  J Autism Dev Disord 2006;36475- 485Google ScholarCrossref
Peele  PBLave  JRKelleher  KJ Exclusions and limitations in children's behavioral health care coverage.  Psychiatr Serv 2002;53591- 594PubMedGoogle ScholarCrossref
 Children's Health Act of 2000, Pub L No. 106-310, 114 Stat 1101 (2000) 
Leslie  DLRosenheck  RAHorwitz  SM Patterns of mental health utilization and costs among children in a privately insured population.  Health Serv Res 2001;36113- 127PubMedGoogle Scholar
Martin  ALeslie  D Psychiatric inpatient, outpatient, and medication utilization and costs among privately insured youths, 1997-2000.  Am J Psychiatry 2003;160757- 764PubMedGoogle ScholarCrossref
Krauss  MWGulley  SSciegaj  MWells  N Access to specialty medical care for children with mental retardation, autism, and other special health care needs.  Ment Retard 2003;41329- 339PubMedGoogle ScholarCrossref