[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Table 1.  
Description of State Autism Spectrum Disorder Insurance Mandates Implemented From 2001 to 2012
Description of State Autism Spectrum Disorder Insurance Mandates Implemented From 2001 to 2012
Table 2.  
Descriptive Statistics for Children in the HCCI Study Sample, 2008-2012
Descriptive Statistics for Children in the HCCI Study Sample, 2008-2012
Table 3.  
Unadjusted and Adjusted Estimates of the Effect of the Autism Spectrum Disorder Insurance Mandates on Treated Prevalence (per 1000)
Unadjusted and Adjusted Estimates of the Effect of the Autism Spectrum Disorder Insurance Mandates on Treated Prevalence (per 1000)
Table 4.  
Adjusted Estimates of the Effect of the Autism Spectrum Disorder Insurance Mandates on Treated Prevalence by Number of Years After Implementation (per 1000)a
Adjusted Estimates of the Effect of the Autism Spectrum Disorder Insurance Mandates on Treated Prevalence by Number of Years After Implementation (per 1000)a
1.
American Psychiatric Association.  Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington, VA: American Psychiatric Association; 2013.
2.
National Autism Center. National Standards Project: phase 2. http://www.nationalautismcenter.org/national-standards-project/phase-2/. Accessed June 3, 2016.
3.
Gillberg  C, Billstedt  E.  Autism and Asperger syndrome: coexistence with other clinical disorders.  Acta Psychiatr Scand. 2000;102(5):321-330.PubMedGoogle ScholarCrossref
4.
Bonde  E.  Comorbidity and subgroups in childhood autism.  Eur Child Adolesc Psychiatry. 2000;9(1):7-10.PubMedGoogle ScholarCrossref
5.
Leyfer  OT, Folstein  SE, Bacalman  S,  et al.  Comorbid psychiatric disorders in children with autism: interview development and rates of disorders.  J Autism Dev Disord. 2006;36(7):849-861.PubMedGoogle ScholarCrossref
6.
Levy  SE, Mandell  DS, Schultz  RT.  Autism.  Lancet. 2009;374(9701):1627-1638.PubMedGoogle ScholarCrossref
7.
Croen  LA, Najjar  DV, Ray  GT, Lotspeich  L, Bernal  P.  A comparison of health care utilization and costs of children with and without autism spectrum disorders in a large group-model health plan.  Pediatrics. 2006;118(4):e1203-e1211.PubMedGoogle ScholarCrossref
8.
Liptak  GS, Stuart  T, Auinger  P.  Health care utilization and expenditures for children with autism: data from U.S. national samples.  J Autism Dev Disord. 2006;36(7):871-879.PubMedGoogle ScholarCrossref
9.
Mandell  DS, Cao  J, Ittenbach  R, Pinto-Martin  J.  Medicaid expenditures for children with autistic spectrum disorders: 1994 to 1999.  J Autism Dev Disord. 2006;36(4):475-485.PubMedGoogle ScholarCrossref
10.
Leslie  DL, Martin  A.  Health care expenditures associated with autism spectrum disorders.  Arch Pediatr Adolesc Med.2007;161(4):350-355.PubMedGoogle ScholarCrossref
11.
Young  A, Ruble  L, McGrew  J.  Public vs private insurance: cost, use, accessibility, and outcomes of services for children with autism spectrum disorders.  Res Autism Spectr Disord. 2009;3(4):1023-1033. doi:10.1016/j.rasd.2009.06.006Google ScholarCrossref
12.
Kogan  MD, Strickland  BB, Blumberg  SJ, Singh  GK, Perrin  JM, van Dyck  PC.  A national profile of the health care experiences and family impact of autism spectrum disorder among children in the United States, 2005-2006.  Pediatrics. 2008;122(6):e1149-e1158.PubMedGoogle ScholarCrossref
13.
Montes  G, Halterman  JS.  Association of childhood autism spectrum disorders and loss of family income.  Pediatrics. 2008;121(4):e821-e826.PubMedGoogle ScholarCrossref
14.
Cidav  Z, Marcus  SC, Mandell  DS.  Implications of childhood autism for parental employment and earnings.  Pediatrics. 2012;129(4):617-623.PubMedGoogle ScholarCrossref
15.
Buescher  AV, Cidav  Z, Knapp  M, Mandell  DS.  Costs of autism spectrum disorders in the United Kingdom and the United States.  JAMA Pediatr. 2014;168(8):721-728.PubMedGoogle ScholarCrossref
16.
Developmental Disabilities Monitoring Network Surveillance Year 2010 Principal Investigators; Centers for Disease Control and Prevention (CDC).  Prevalence of autism spectrum disorder among children aged 8 years—Autism and Developmental Disabilities Monitoring Network, 11 sites, United States, 2010.  MMWR Surveill Summ. 2014;63(2):1-21.PubMedGoogle Scholar
17.
Bouder  JN, Spielman  S, Mandell  DS.  Brief report: quantifying the impact of autism coverage on private insurance premiums.  J Autism Dev Disord. 2009;39(6):953-957.PubMedGoogle ScholarCrossref
18.
Newschaffer  CJ, Curran  LK.  Autism: an emerging public health problem.  Public Health Rep. 2003;118(5):393-399.PubMedGoogle ScholarCrossref
19.
Shattuck  PT, Grosse  SD.  Issues related to the diagnosis and treatment of autism spectrum disorders.  Ment Retard Dev Disabil Res Rev. 2007;13(2):129-135.PubMedGoogle ScholarCrossref
20.
Baller  JB, Barry  CL, Shea  K, Walker  MM, Ouellette  R, Mandell  DS.  Assessing early implementation of state autism insurance mandates [published online November 27, 2015].  Autism.PubMedGoogle Scholar
21.
Rogers  SJ, Vismara  LA.  Evidence-based comprehensive treatments for early autism.  J Clin Child Adolesc Psychol. 2008;37(1):8-38.PubMedGoogle ScholarCrossref
22.
Barry  CL, Gabel  JR, Frank  RG, Hawkins  S, Whitmore  HH, Pickreign  JD.  Design of mental health benefits: still unequal after all these years.  Health Aff (Millwood). 2003;22(5):127-137.PubMedGoogle ScholarCrossref
23.
Bilaver  LA, Jordan  N.  Impact of state mental health parity laws on access to autism services.  Psychiatr Serv. 2013;64(10):967-973.PubMedGoogle ScholarCrossref
24.
Parish  S, Thomas  K, Rose  R, Kilany  M, McConville  R.  State insurance parity legislation for autism services and family financial burden.  Intellect Dev Disabil. 2012;50(3):190-198.PubMedGoogle ScholarCrossref
25.
Autism Speaks. State initiatives. http://www.autismspeaks.org/state-initiatives. Accessed June 3, 2016.
26.
Employee Retirement Income Security Act. Pub L 93–406, 88 Stat 829. Enacted September 2, 1974, codified in part at 29 USC ch 18.
27.
Buchmueller  TC, Cooper  PF, Jacobson  M, Zuvekas  SH.  Parity for whom? exemptions and the extent of state mental health parity legislation.  Health Aff (Millwood). 2007;26(4):w483-w487.PubMedGoogle ScholarCrossref
28.
Wang  L, Mandell  DS, Lawer  L, Cidav  Z, Leslie  DL.  Healthcare service use and costs for autism spectrum disorder: a comparison between Medicaid and private insurance.  J Autism Dev Disord. 2013;43(5):1057-1064.PubMedGoogle ScholarCrossref
29.
Stein  BD, Sorbero  MJ, Goswami  U, Schuster  J, Leslie  DL.  Impact of a private health insurance mandate on public sector autism service use in Pennsylvania.  J Am Acad Child Adolesc Psychiatry. 2012;51(8):771-779.PubMedGoogle ScholarCrossref
30.
Wang  L, Leslie  DL.  Health care expenditures for children with autism spectrum disorders in Medicaid.  J Am Acad Child Adolesc Psychiatry. 2010;49(11):1165-1171.PubMedGoogle Scholar
31.
Burke  JP, Jain  A, Yang  W,  et al.  Does a claims diagnosis of autism mean a true case?  Autism. 2014;18(3):321-330.PubMedGoogle ScholarCrossref
32.
Mandell  DS, Morales  KH, Xie  M, Polsky  D, Stahmer  A, Marcus  SC.  County-level variation in the prevalence of Medicaid-enrolled children with autism spectrum disorders.  J Autism Dev Disord. 2010;40(10):1241-1246.PubMedGoogle ScholarCrossref
33.
Chatterji  P, Decker  SL, Markowitz  S.  The effects of mandated health insurance benefits for autism on out-of-pocket costs and access to treatment.  J Policy Anal Manage. 2015;34(2):328-353.PubMedGoogle ScholarCrossref
34.
Gruber  J.  State-mandated benefits and employer-provided health insurance.  J Public Econ. 1994;55(3):433-464. doi:10.1016/0047-2727(93)01407-2Google ScholarCrossref
35.
Schmidt  L.  Effects of infertility insurance mandates on fertility.  J Health Econ. 2007;26(3):431-446.PubMedGoogle ScholarCrossref
36.
Allison  P.  Logistic Regression Using the SAS System: Theory and Application. Cary, NC: John Wiley and Sons, Inc; 1999.
37.
Cameron  AC, Miller  DL.  A practitioner’s guide to cluster-robust inference.  J Hum Resour. 2015;50(2):317-372. doi:10.3368/jhr.50.2.317Google ScholarCrossref
Original Investigation
September 2016

Effects of Autism Spectrum Disorder Insurance Mandates on the Treated Prevalence of Autism Spectrum Disorder

Author Affiliations
  • 1Center for Mental Health Policy and Services Research, University of Pennsylvania Perelman School of Medicine, Philadelphia
  • 2Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia
  • 3Department of Health Policy and Management, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland
  • 4University of Pennsylvania School of Social Policy and Practice, Philadelphia
  • 5Division of General Internal Medicine, University of Pennsylvania Perelman School of Medicine, Philadelphia
JAMA Pediatr. 2016;170(9):887-893. doi:10.1001/jamapediatrics.2016.1049
Abstract

Importance  Most states have passed insurance mandates requiring commercial health plans to cover services for children with autism spectrum disorder (ASD). Insurers have expressed concerns that these mandates will increase the number of children diagnosed with ASD (treated prevalence) and therefore increase costs associated with their care. To our knowledge, no published studies have addressed this question.

Objective  To examine whether implementing ASD insurance mandates increases the number of commercially insured children diagnosed with ASD.

Design, Setting, and Participants  A difference-in-differences study was performed using inpatient and outpatient health insurance claims for children 21 years or younger covered by 3 of the largest insurers in the United States—United HealthCare, Aetna, and Humana—from January 1, 2008, through December 31, 2012, made available through the Health Care Cost Institute. Data analysis was conducted from March 15 to August 11, 2015.

Exposures  Implementation of an ASD insurance mandate in a child’s state of residence.

Main Outcomes and Measures  The treated prevalence of ASD, measured as a binary indicator of whether a given child in a given calendar month had at least 1 health care service claim associated with a diagnosis of ASD.

Results  The adjusted treated prevalence among 1 046 850 eligible children (575 299 male [55.0%]) in states with ASD insurance mandates was 1.8 per 1000 and 1.6 per 1000 among children in states without such a mandate (P = .006). The mean increase in treated prevalence attributable to the mandates was 0.21 per 1000 children during the study period (95% CI, 0.11-0.30; P < .001). Mandates in place longer had a larger effect on treated prevalence. The mean increase in treated prevalence of ASD attributable to the mandate was 0.17 per 1000 children (95% CI, 0.09-0.24; P < .001) in the first year following implementation, 0.27 per 1000 children (95% CI, 0.13-0.42; P < .001) in the second year, and 0.29 per 1000 children (95% CI, 0.15-0.42; P < .001) 3 years or more following implementation.

Conclusions and Relevance  Implementing state ASD insurance mandates resulted in increases in the number of children diagnosed with ASD; these numbers increased each year after implementation. Even 3 years or more after implementation, however, treated prevalence of ASD was much lower than community prevalence estimates. This finding may allay concerns that mandates will substantially increase insurance costs, but it suggests that many commercially insured children with ASD remain undiagnosed or are being treated only through publicly funded systems.

Introduction

Autism spectrum disorder (ASD) is characterized by impairments in social communication, highly restricted interests, and repetitive behaviors.1 Recommended treatments include behavioral and educational interventions delivered for as many as 25 hours per week for several years.2 Autism spectrum disorder is often accompanied by other health conditions that require additional treatment.3-6 As a result, cost of health care7-10 and related costs11-15 are much higher for children with ASD than for typically developing children.

The Centers for Disease Control and Prevention estimates that 1 in 68 children aged 8 years in the United States met the criteria for ASD in 2010, an increase from 1 in 150 children in 2004.16 The intensive health care needs and growing numbers of children with ASD have resulted in considerable debate regarding how best to pay for their care.17-19

Until recently, commercial health insurance plans typically did not cover treatments for children with ASD,17,20 usually on the grounds that they were unproven or experimental.21 Even when treatments for ASD were covered under behavioral health benefits, their recommended intensity far outpaced the typically covered number of visits.22 Other policy mechanisms, such as mental health parity laws, have met with mixed success in addressing the needs of children with ASD.23,24 In response, 42 states have passed ASD insurance mandates.25 These laws require commercial, fully insured plans to cover ASD-specific behavioral therapies, with annual caps ranging from $12 000 to $50 000 depending on the state and age of the child. Despite some differences, the laws’ intent was to make behavioral treatments for ASD a mandatory part of commercial health insurance benefits.

State insurance mandates apply to a subset of a state’s privately insured population. The Employee Retirement Income Security Act26 exempts self-insured firms—those that contract with health plans only to administer employee health benefits and not to pool risk—from state insurance regulations. About half of privately insured individuals are covered through these plans27 and therefore are not affected by the ASD insurance mandates.

Insurance companies fought the mandates, arguing that the number of enrollees diagnosed with ASD would increase and that they would use services up to the annual dollar caps, resulting in drastic increases in spending.17 Indeed, the treated prevalence of ASD—that is, the number of individuals diagnosed with ASD in the health care system—has been far below the community prevalence estimated by the Centers for Disease Control and Prevention. Studies using insurance claims found that between 2 and 7 children per 1000 received a diagnosis of ASD on a claim.7,28-32 Increasing the number of commercially insured children diagnosed with ASD so that it mirrored the community prevalence would constitute a 2- to 5-fold increase.

Advocates for individuals with ASD have celebrated these mandates, but little evidence exists on their effects. Chatterji et al33 found that the mandates had no effect on costs or access to treatment among privately insured children. Their data did not allow them to determine whether a child was enrolled in an insurance plan subject to the mandate, however, nor did they examine the effects of the mandates on treated prevalence. Using data from 1 state, Stein et al29 found that the mandate reduced the growth in children with ASD using publicly funded services. They did not examine whether this change resulted in a parallel growth in children with ASD receiving care reimbursed through private insurance.

To address the question of whether state ASD insurance mandates have increased treated prevalence, we used pooled data from 3 large, national health insurance companies. We used a difference-in-differences approach that exploited variation in the implementation of mandates across states and during the 5-year study period and compared eligible and ineligible children within states. We also examined whether the mandates’ effects varied with time since their implementation, as policy changes often have lagged effects. A recent study found, for example, that full implementation of a mandate can be delayed by lack of regulatory guidance regarding credentialing and limited capacity of health care professionals.20 Results from our study can help states that are considering enacting ASD insurance mandates25 and states with mandates that are planning for their effects.

Box Section Ref ID

Key Points

  • Question What is the effect of autism spectrum disorder (ASD) insurance mandates on the treated prevalence of ASD?

  • Findings In the first year after mandates were implemented, the treated prevalence of ASD increased 10% relative to states without mandates, with additional increases in the second and later years. Even 3 years or more after implementation, however, the treated prevalence was only 1.8 per 1000 children.

  • Meaning Autism spectrum disorder insurance mandates resulted in increases in the number of children diagnosed with ASD through private insurance, but many commercially insured children with ASD remain undiagnosed.

Methods
Data Sources and Sample

To examine the treated prevalence of ASD before and after mandates were implemented, we used inpatient and outpatient facility and professional procedure claims data from the Health Care Cost Institute from January 1, 2008, through December 31, 2012. Data were provided from 3 of the nation’s largest insurers: United HealthCare, Aetna, and Humana. These commercial claims represent more than 50 million individuals per year in all 50 states and the District of Columbia. An important aspect of the Health Care Cost Institute data is that they indicate whether an enrollee had commercial insurance coverage through a firm that was self-insured or fully insured.

Our sample included children from birth through 21 years. We identified 154 089 children diagnosed with ASD on the basis of a claim associated with an International Classification of Diseases, Ninth Revision, Clinical Modification, code of 299.xx any time during the 5-year study period. Child–calendar month was the unit of analysis. To be sure that we had all information on service use, we included only enrollees with commercial insurance for whom mental health claims were available. We also excluded individually insured plans, which represented less than 0.2% of all child-months in the data set. The University of Pennsylvania Institutional Review Board approved this study. All patient data were deidentified.

Variables

The study outcome was a binary indicator of whether a given child in a given calendar month had at least 1 claim for a health care service associated with a diagnosis of ASD. Our 2 key binary independent variables were whether a state had an active mandate at a given time point and whether a child would have been subject to an ASD mandate if his or her state had one (as determined by enrollment in a fully insured plan and the child’s age). In states with mandates, we limited age to that specified in the mandate (eTable 1 in the Supplement).

To identify whether a child resided in a state with an implemented ASD mandate law in a specific month, we compiled information from Autism Speaks (https://www.autismspeaks.org/state-initiatives) detailing which states enacted mandate laws, the month of enactment, and the relevant provisions, including to what age range they applied. We then verified this information by reviewing the original mandate laws (http://www.ncsl.org/research/health/autism-legislation-database.aspx). Mandates included in this study are described in Table 1.

In addition, we created a binary indicator for whether a specific child in a given month would have been subject to an ASD insurance mandate if his or her state had one. This indicator was based on whether the child had employer-based insurance that was fully insured, and, for states that implemented mandates, the mandate’s criteria for the child’s age. For states that did not pass a mandate during the study period, we included individuals 21 years or younger because that is the modal age range covered under the mandates (10 states). In a sensitivity analysis, we tested the effects of the mandates in a sample in which eligibility in states without an ASD insurance mandate was limited to those 18 years or younger (eAppendix in the Supplement). Other variables besides state and calendar year included the following: child sex, age in the given month (estimated based on July 1 in their year of birth), insurance product type (health maintenance organization, point of service, preferred provider organization, exclusive provider organization, or indemnity or other), whether the child was in a high-deductible plan, and calendar month.

Statistical Analysis

Data analysis was conducted from March 15 to August 11, 2015. To evaluate the effect of state ASD laws on treated prevalence, we used a difference-in-differences approach with state and year fixed effects34,35 that compared treated prevalence within states before and after mandate implementation and between groups of children who would and would not be affected by the laws based on the source of their health insurance and their age. Our treatment group consisted of children who lived in states with active ASD insurance mandates and who were eligible for the mandate. Our 3 comparison groups consisted of children in states with an active mandate who were not subject to the mandate, children in states without a mandate who would have been subject to the mandate if one were active, and children in states without a mandate who would not have been subject to the mandate if one were active. Our use of comparison groups of children in states without mandates accounted for secular trends in treated prevalence unrelated to state ASD mandate laws.

Because of limited computing resources, it was not feasible to analyze the full data set, which consisted of more than 600 million child-month observations. Instead, we followed an approach described by Allison36 that combined stratified random sampling on the dependent variable and logistic regression. By including in our analytic cohort all child-month observations with an ASD diagnosis and an equal number of randomly selected child-months without an ASD diagnosis, logistic regression on this cohort yields unbiased coefficient estimates. The intercept is biased in a known fashion and was corrected manually after estimation.

We calculated descriptive statistics for the overall sample of children and those in each of the 4 groups. Next, we estimated logistic regression models on the child-month–level binary outcome of treated prevalence. Our simplest unadjusted difference-in-differences model specification included only indicators for whether a child-month was located in a state with an active ASD insurance mandate, whether the child-month was eligible for a mandate if one were active, and their interaction. The adjusted analysis added child-level controls and state and year fixed effects. In addition, we estimated a second set of unadjusted and adjusted logistic regression models that included 3 cohort indicators to examine whether the effects of the state mandates differed based on how many years a law had been in place. Standard errors were adjusted to account for the clustering of observations within states.37 To ease interpretation, we converted the model results to predictive margins on the probability scale and used Wald tests to compare predictive margins statistically. Attempts to perform statistical inference by means of a percentile bootstrap were not successful with the limited available computing resources.

To bolster our confidence in our quasi-experimental approach, we compared trends in treated prevalence in the years before implementation of the ASD insurance mandate among states that eventually did and did not adopt such a mandate. We also conducted several sensitivity analyses, including testing our assumptions regarding the ages of children that would have been eligible for the mandates had states without mandates enacted them, the potential for the mandates to have differential effects on different age groups, and the inclusion in our analysis of states that already had enacted a mandate at the beginning of the study period. The eAppendix in the Supplement includes more information about our analytic approach, including estimating equations.

Results

Table 1 presents the characteristics of the 29 state mandates implemented during the study period. Indiana was the first state to implement an ASD insurance mandate in 2001, followed by 3 states in 2008 (Illinois, South Carolina, and Texas), before the start of the study period. During our study period, mandates were implemented by 6 states in 2009 (Arizona, Florida, Louisiana, New Mexico, Pennsylvania, and Wisconsin), 4 states in 2010 (Colorado, Connecticut, Montana, and New Jersey), 8 states in 2011 (Arkansas, Kentucky, Massachusetts, Maine, Missouri, New Hampshire, Nevada, and Vermont), and 7 states in 2012 (California, Delaware, Michigan, New York, Rhode Island, Virginia, and West Virginia). Two states (Texas and Vermont) expanded the age range covered 2 years after initial implementation. All state mandates applied to fully insured firms with more than 50 employees; 22 of the 29 state mandates also applied to fully insured firms with 50 or fewer employees. By the end of the study period, 18 of the 29 mandates covered individuals from birth through 18 years or older.

Table 2 describes the sample. In all 4 groups, approximately 55% of the sample was male. The mean age on study entry was between 10 and 11 years, and the modal age range was 6 to 12 years (approximately one-third). The percentage of youth between the ages of 18 and 21 years differed notably among the 4 groups, ranging from 14.8% to 20.0%, but other age groups were roughly equivalent across the groups. As expected, insurance product type differed considerably by group, with the groups for the ASD insurance mandate less likely to be enrolled in health maintenance organizations and more likely to be enrolled in point-of-service plans. The unadjusted prevalence of ASD in the sample was 1.7 per 1000 children. It was higher among males than females and among children aged 6 to 12 years than other age groups. The prevalence was similar across insurance types.

Table 3 presents the results from the unadjusted and adjusted models examining the effect of the state ASD insurance mandates on treated prevalence. The unadjusted treated prevalence among eligible children in states with an ASD insurance mandate was 2.0 per 1000 children compared with 1.4 per 1000 children in states without a mandate (P = .002). The unadjusted difference-in-differences estimate was 0.44 per 1000 children (95% CI, –0.002 to 0.88; P = .05). In the adjusted model, the treated prevalence among eligible children in states with an ASD insurance mandate was 1.8 per 1000 children compared with 1.6 per 1000 children in states without a mandate (P = .006). The adjusted difference-in-differences estimate was 0.21 per 1000 children (95% CI, 0.11-0.30; P < .001). Measured in proportional terms, the mandates were associated with a 12.7% adjusted increase in treated prevalence relative to the 1.6 per 1000 among eligible children in states without a mandate.

We also assessed whether the magnitude of the effect of state ASD mandate laws on treated prevalence differed by the number of years a mandate had been implemented (Table 4). Compared with the treated prevalence rate of 1.6 per 1000 among eligible children in states without mandates, the treated prevalence rates among eligible children in states with an ASD mandate in the first, second, and third or more years after implementation were 1.7, 1.8, and 1.8 per 1000 children, respectively. These adjusted differences in treated prevalence of 0.17 (95% CI, 0.09-0.24; P < .001), 0.27 (95% CI, 0.13-0.42; P < .001), and 0.29 (95% CI, 0.15-0.42; P < .001) per 1000 attributable to the mandates were associated with a 10.4% increase in treated prevalence in the first year of implementation, a 17.1% increase in the second year, and an 18.0% increase in the third and later years after implementation relative to the 1.6 per 1000 baseline among eligible children in states without a mandate.

We saw no evidence of differential trends in the treated prevalence between eligible and ineligible children in states that never implemented an ASD insurance mandate compared with states that eventually did (eTable 2 in the Supplement). As expected, there were no statistically significant differences in treated prevalence between eligible and ineligible children in states that never implemented mandates. Treated prevalence was not statistically significantly different between states that never implemented mandates and the pre-implementation period among states that eventually implemented mandates. Likewise, the difference in treated prevalence between eligible and ineligible children was not statistically significantly different between these 2 groups of states. There were no appreciable effects of children’s age on the effects of the mandate. In states that eventually implemented a mandate, the trends in treated prevalence before implementation did not differ statistically significantly between eligible and ineligible children (eAppendix in the Supplement).

Discussion

In this study, implementing ASD insurance mandates was associated with an increase of about 10% in the treated prevalence of ASD among eligible children, and the increase attributable to the mandates rose to 18% after the mandates had been in place for a few years. This is an encouraging finding, suggesting that the policies are at least partially meeting their goal of increasing the number of children with ASD who receive care reimbursed through private insurance; however, the monthly treated prevalence of 1.8 per 1000 children 3 years after implementation of the mandates is still well below what one might expect had all privately insured children received care reimbursed by their insurer. Although it is perhaps not a fair direct comparison, the Centers for Disease Control and Prevention’s estimate of the community prevalence of ASD is 15 per 1000 children.16 Given that the recommended ASD treatment intensity would indicate monthly—if not weekly—visits for behavioral, occupational, and speech therapies, our results suggest that children receiving that intensity of service are obtaining it through schools and public insurance, not private insurance.

Several factors may be responsible for the relatively modest increase in treated prevalence in the first year of the ASD insurance mandate, followed by additional increases in subsequent years. Our interviews with stakeholders suggest that the regulatory and credentialing processes often were vague and difficult for health care professionals to comply with, reimbursements for assessment and treatment were low, and there often was a dearth of well-trained clinicians to meet the growing demand for care.20 Insurers and clinicians may have required additional time to develop these processes, to increase the supply of clinicians, and for families to begin to access the ASD insurance benefit.

Several study limitations should be mentioned. First, diagnoses of ASD in claims were not confirmed through observation or clinical interview. A prior study using similar data found that the criterion used in our study of 1 claim associated with an ASD diagnosis yielded a positive predictive value of 87.4%, suggesting few false-positives.31 Second, our study did not account for differences in the implementation of the mandate by state beyond age requirements. Some states have more stringent dollar caps or may have more aggressively required insurance companies to identify and serve children with ASD, resulting in greater treated prevalence. Third, while our data set contains a substantial portion of privately insured children in the United States, factors specific to these insurance plans may limit generalizability of the findings. Our observed treated prevalence was about equal to the median of prior studies using claims data,7,28-32 suggesting that our finding is not an artifact of our data source.

Conclusions

Despite these limitations, there are important implications related to these findings. Mandates have had a promising effect on increasing the number of commercially insured children diagnosed with ASD and the effect increases 2 years after implementation; however, that number is still well below the community prevalence of ASD. On the one hand, this finding should allay insurers’ concerns regarding potential sizable increases in cost.17 On the other hand, the mandates have not had the full effect that advocates desired. The results suggest the need for additional strategies to enforce the mandates and address barriers, such as regulatory issues or clinician capacity, that inhibit the timely and appropriate identification of children with ASD.20

Back to top
Article Information

Accepted for Publication: April 8, 2016.

Corresponding Author: David S. Mandell, ScD, Center for Mental Health Policy and Services Research, University of Pennsylvania Perelman School of Medicine, 3535 Market St, 3rd Floor, Philadelphia, PA 19104 (mandelld@upenn.edu).

Published Online: July 11, 2016. doi:10.1001/jamapediatrics.2016.1049.

Author Contributions: Drs Mandell and Epstein had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Mandell, Barry, Marcus, Epstein.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Mandell, Barry, Xie, Shea, Mullan.

Critical revision of the manuscript for important intellectual content: Mandell, Barry, Marcus, Epstein.

Statistical analysis: Marcus, Xie, Epstein.

Obtained funding: Mandell, Barry.

Administrative, technical, or material support: Mandell, Barry, Shea, Mullan, Epstein.

Study supervision: Mandell, Barry.

Conflict of Interest Disclosures: Dr Marcus reported receiving consulting fees from Alkermes, Shire, Johnson and Johnson, Forest, and Sunovion. No other disclosures were reported.

References
1.
American Psychiatric Association.  Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington, VA: American Psychiatric Association; 2013.
2.
National Autism Center. National Standards Project: phase 2. http://www.nationalautismcenter.org/national-standards-project/phase-2/. Accessed June 3, 2016.
3.
Gillberg  C, Billstedt  E.  Autism and Asperger syndrome: coexistence with other clinical disorders.  Acta Psychiatr Scand. 2000;102(5):321-330.PubMedGoogle ScholarCrossref
4.
Bonde  E.  Comorbidity and subgroups in childhood autism.  Eur Child Adolesc Psychiatry. 2000;9(1):7-10.PubMedGoogle ScholarCrossref
5.
Leyfer  OT, Folstein  SE, Bacalman  S,  et al.  Comorbid psychiatric disorders in children with autism: interview development and rates of disorders.  J Autism Dev Disord. 2006;36(7):849-861.PubMedGoogle ScholarCrossref
6.
Levy  SE, Mandell  DS, Schultz  RT.  Autism.  Lancet. 2009;374(9701):1627-1638.PubMedGoogle ScholarCrossref
7.
Croen  LA, Najjar  DV, Ray  GT, Lotspeich  L, Bernal  P.  A comparison of health care utilization and costs of children with and without autism spectrum disorders in a large group-model health plan.  Pediatrics. 2006;118(4):e1203-e1211.PubMedGoogle ScholarCrossref
8.
Liptak  GS, Stuart  T, Auinger  P.  Health care utilization and expenditures for children with autism: data from U.S. national samples.  J Autism Dev Disord. 2006;36(7):871-879.PubMedGoogle ScholarCrossref
9.
Mandell  DS, Cao  J, Ittenbach  R, Pinto-Martin  J.  Medicaid expenditures for children with autistic spectrum disorders: 1994 to 1999.  J Autism Dev Disord. 2006;36(4):475-485.PubMedGoogle ScholarCrossref
10.
Leslie  DL, Martin  A.  Health care expenditures associated with autism spectrum disorders.  Arch Pediatr Adolesc Med.2007;161(4):350-355.PubMedGoogle ScholarCrossref
11.
Young  A, Ruble  L, McGrew  J.  Public vs private insurance: cost, use, accessibility, and outcomes of services for children with autism spectrum disorders.  Res Autism Spectr Disord. 2009;3(4):1023-1033. doi:10.1016/j.rasd.2009.06.006Google ScholarCrossref
12.
Kogan  MD, Strickland  BB, Blumberg  SJ, Singh  GK, Perrin  JM, van Dyck  PC.  A national profile of the health care experiences and family impact of autism spectrum disorder among children in the United States, 2005-2006.  Pediatrics. 2008;122(6):e1149-e1158.PubMedGoogle ScholarCrossref
13.
Montes  G, Halterman  JS.  Association of childhood autism spectrum disorders and loss of family income.  Pediatrics. 2008;121(4):e821-e826.PubMedGoogle ScholarCrossref
14.
Cidav  Z, Marcus  SC, Mandell  DS.  Implications of childhood autism for parental employment and earnings.  Pediatrics. 2012;129(4):617-623.PubMedGoogle ScholarCrossref
15.
Buescher  AV, Cidav  Z, Knapp  M, Mandell  DS.  Costs of autism spectrum disorders in the United Kingdom and the United States.  JAMA Pediatr. 2014;168(8):721-728.PubMedGoogle ScholarCrossref
16.
Developmental Disabilities Monitoring Network Surveillance Year 2010 Principal Investigators; Centers for Disease Control and Prevention (CDC).  Prevalence of autism spectrum disorder among children aged 8 years—Autism and Developmental Disabilities Monitoring Network, 11 sites, United States, 2010.  MMWR Surveill Summ. 2014;63(2):1-21.PubMedGoogle Scholar
17.
Bouder  JN, Spielman  S, Mandell  DS.  Brief report: quantifying the impact of autism coverage on private insurance premiums.  J Autism Dev Disord. 2009;39(6):953-957.PubMedGoogle ScholarCrossref
18.
Newschaffer  CJ, Curran  LK.  Autism: an emerging public health problem.  Public Health Rep. 2003;118(5):393-399.PubMedGoogle ScholarCrossref
19.
Shattuck  PT, Grosse  SD.  Issues related to the diagnosis and treatment of autism spectrum disorders.  Ment Retard Dev Disabil Res Rev. 2007;13(2):129-135.PubMedGoogle ScholarCrossref
20.
Baller  JB, Barry  CL, Shea  K, Walker  MM, Ouellette  R, Mandell  DS.  Assessing early implementation of state autism insurance mandates [published online November 27, 2015].  Autism.PubMedGoogle Scholar
21.
Rogers  SJ, Vismara  LA.  Evidence-based comprehensive treatments for early autism.  J Clin Child Adolesc Psychol. 2008;37(1):8-38.PubMedGoogle ScholarCrossref
22.
Barry  CL, Gabel  JR, Frank  RG, Hawkins  S, Whitmore  HH, Pickreign  JD.  Design of mental health benefits: still unequal after all these years.  Health Aff (Millwood). 2003;22(5):127-137.PubMedGoogle ScholarCrossref
23.
Bilaver  LA, Jordan  N.  Impact of state mental health parity laws on access to autism services.  Psychiatr Serv. 2013;64(10):967-973.PubMedGoogle ScholarCrossref
24.
Parish  S, Thomas  K, Rose  R, Kilany  M, McConville  R.  State insurance parity legislation for autism services and family financial burden.  Intellect Dev Disabil. 2012;50(3):190-198.PubMedGoogle ScholarCrossref
25.
Autism Speaks. State initiatives. http://www.autismspeaks.org/state-initiatives. Accessed June 3, 2016.
26.
Employee Retirement Income Security Act. Pub L 93–406, 88 Stat 829. Enacted September 2, 1974, codified in part at 29 USC ch 18.
27.
Buchmueller  TC, Cooper  PF, Jacobson  M, Zuvekas  SH.  Parity for whom? exemptions and the extent of state mental health parity legislation.  Health Aff (Millwood). 2007;26(4):w483-w487.PubMedGoogle ScholarCrossref
28.
Wang  L, Mandell  DS, Lawer  L, Cidav  Z, Leslie  DL.  Healthcare service use and costs for autism spectrum disorder: a comparison between Medicaid and private insurance.  J Autism Dev Disord. 2013;43(5):1057-1064.PubMedGoogle ScholarCrossref
29.
Stein  BD, Sorbero  MJ, Goswami  U, Schuster  J, Leslie  DL.  Impact of a private health insurance mandate on public sector autism service use in Pennsylvania.  J Am Acad Child Adolesc Psychiatry. 2012;51(8):771-779.PubMedGoogle ScholarCrossref
30.
Wang  L, Leslie  DL.  Health care expenditures for children with autism spectrum disorders in Medicaid.  J Am Acad Child Adolesc Psychiatry. 2010;49(11):1165-1171.PubMedGoogle Scholar
31.
Burke  JP, Jain  A, Yang  W,  et al.  Does a claims diagnosis of autism mean a true case?  Autism. 2014;18(3):321-330.PubMedGoogle ScholarCrossref
32.
Mandell  DS, Morales  KH, Xie  M, Polsky  D, Stahmer  A, Marcus  SC.  County-level variation in the prevalence of Medicaid-enrolled children with autism spectrum disorders.  J Autism Dev Disord. 2010;40(10):1241-1246.PubMedGoogle ScholarCrossref
33.
Chatterji  P, Decker  SL, Markowitz  S.  The effects of mandated health insurance benefits for autism on out-of-pocket costs and access to treatment.  J Policy Anal Manage. 2015;34(2):328-353.PubMedGoogle ScholarCrossref
34.
Gruber  J.  State-mandated benefits and employer-provided health insurance.  J Public Econ. 1994;55(3):433-464. doi:10.1016/0047-2727(93)01407-2Google ScholarCrossref
35.
Schmidt  L.  Effects of infertility insurance mandates on fertility.  J Health Econ. 2007;26(3):431-446.PubMedGoogle ScholarCrossref
36.
Allison  P.  Logistic Regression Using the SAS System: Theory and Application. Cary, NC: John Wiley and Sons, Inc; 1999.
37.
Cameron  AC, Miller  DL.  A practitioner’s guide to cluster-robust inference.  J Hum Resour. 2015;50(2):317-372. doi:10.3368/jhr.50.2.317Google ScholarCrossref
×