Key Points
Question
What is the prevalence of identified autism in Medicaid-enrolled adults, and how has Medicaid enrollment changed over time and by other demographic factors?
Findings
In this cohort study of 403 028 Medicaid enrollees with autism claims over 9 years of claims data, autism prevalence increased from 4.2 per 1000 enrollees in 2011 to 9.5 per 1000 enrollees in 2019. The largest increase was observed in the 25- to 34-year age group and the smallest increase in the 55- to 64-year age group.
Meaning
These findings suggest that despite difficulties in identifying autism in adults, there is a considerable and growing population of autistic adults among Medicaid enrollees, which may have future implications for the Medicaid system and understanding the needs of the autistic population.
This cohort study uses Medicaid claims data to estimate the prevalence of autistic adult enrollees.
Abstract
Importance
The reported prevalence of autism in children has consistently risen over the past 20 years. The concurrent implications for the adult Medicaid system, which insures autistic adults due to low income or disability, have not been studied.
Objective
To estimate the prevalence of adults identified as autistic in Medicaid claims data and to examine the prevalence by year, age, and race and ethnicity to understand enrollment patterns.
Design, Setting, and Participants
This cohort study used data from a longitudinal Medicaid claims cohort of enrollees aged 18 years or older with a claim for autism at any point from January 1, 2011, to December 31, 2019, and an approximately 1% random sample of all adult Medicaid enrollees. The data were analyzed between February 22 and June 22, 2023.
Exposure
Adults enrolled in Medicaid with a claim for autism.
Main Outcome and Measures
Prevalence of autism per 1000 Medicaid enrollees for each year was calculated using denominator data from the Centers for Medicare & Medicaid Services weighted to nondisabled population demographic characteristics. Prevalence by race and ethnicity were calculated for study year 2019.
Results
Across 9 years, 403 028 unique adults had autism claims in their Medicaid records (25.7% female, 74.2% male, 3.3% Asian, 16.8% Black, 12.2% Hispanic, 0.8% Native American, 0.8% Pacific Islander, 74.3% White, and 4.2% of multiple races). Across all ages, autism prevalence increased from 4.2 per 1000 enrollees in 2011 to 9.5 per 1000 enrollees in 2019. The largest increase over the 9 years was in the 25- to 34-year age group (195%), and the smallest increase was in the 55- to 64-year age group (45%). The prevalence of White enrollees was at least 2 times that of the prevalence of every other racial group in all age categories.
Conclusions and Relevance
The study findings suggest that despite difficulties in identifying autism in adults, there is a considerable and growing population of autistic adults enrolled in Medicaid. As children on the autism spectrum become autistic adults, Medicaid is an important insurance provider for an increasing number of autistic adults and can be a valuable resource for understanding the health of the autistic population.
Introduction
The prevalence of autism in children has risen over the past 20 years.1 While the rise may be attributed to improved diagnostic practice, genetics, and environmental exposures,2 each year, more children need autism-related services.3 The need for autism services, including health care services, does not end at childhood.4
As the prevalence of children identified as autistic has increased, there is a concomitant increase of autistic adults in the Medicaid system.5,6 Medicaid is a crucial insurance provider for autistic adults.5 Because of the impairment associated with autism and its co-occurring conditions and persistent stigma, exclusion, and ableism, many autistic adults are not employed full-time7; thus, many are eligible for and rely on Medicaid.
Historically, autistic people from marginalized groups have lower access to health care. Those from minoritized racial and ethnic groups have been diagnosed less and later and have received fewer services than White peers.8,9,10 Some of this disparity has decreased over time, specifically autism ascertainment in schools,1,2 but still exists across the life course. These diagnostic disparities may be present in Medicaid but have not been quantified. Our goal was to calculate the prevalence of adults identified as autistic in Medicaid claims data and understand enrollment patterns by year and enrollees’ age and race and ethnicity.
Methods
The data used in this cohort study are from a longitudinal cohort of all adults 18 years or older enrolled in Medicaid in the US between January 1, 2011, and December 31, 2019; a subset of those with 1 or more fee-for-service or managed care claims with an International Classification of Diseases (ICD) code of autism (any version); and a random sample of other enrollees. The Boston University Medical Campus institutional review board deemed this study not human participants research and waived informed consent. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
We used established algorithms to determine autism case status (eTable in Supplement 1) from inpatient, long-term-care, or other service claims files. Individuals meeting autism criteria in any year were considered part of the autism group in all years they were enrolled, independent of whether they had claims in that year. We aligned 2 Medicaid data systems to create the longitudinal sample: Medicaid Analytic eXtract (2011-2015) and Transformed Medicaid Statistical Information System Analytic Files (2014-2019).
Age was determined by date of birth and categorized into 6 age groups. Race and ethnicity were self-reported using the state’s specified options, then reclassified as Asian, Black, Native American, Pacific Islander, White, and multiple races (≥2 races listed across years) by Centers for Medicare & Medicaid Services (CMS) criteria. Race or ethnicity reported in any year was considered that individual’s race or ethnicity, even if other years had missing information. We used multiple imputation models to account for missing race and ethnicity data11 (eMethods in Supplement 1). Ethnicity data were collected dichotomously (Hispanic, not Hispanic) and, thus, were analyzed dichotomously because no more-specific Hispanic ethnicities were available in the data.
The data analysis was performed between February 22 and June 22, 2023. We calculated demographic statistics for 9 years and Medicaid prevalence using total enrolled population denominator data from CMS data books.12 We weighted the full denominator by the age and race and ethnicity distribution in our approximately 1% sample of random Medicaid enrollees to calculate age- and race and ethnicity–specific prevalence. Because of the large sample size, we do not present 95% CIs, since intervals are too precise to be meaningful. We conducted quantitative bias analysis to evaluate outcomes of misclassification due to imperfect positive and negative predictive values of autism claims13 (eMethods in Supplement 1). Statistical analyses were performed using R, version 4.3.0 (R Foundation for Statistical Computing).
Results
Across 9 years, 403 028 adults with Medicaid had autism claims (Table). Demographically, 16.8% of autistic adults were Black and 12.2% were Hispanic (vs 3.3% Asian, 0.8% Native American, 0.8% Pacific Islander, 74.2% White, and 4.2% of multiple races). Compared with our random Medicaid sample, autistic adults were younger and more likely to be non-Hispanic White. Documented sex of autistic adults was 25.7% female and 74.3% male, which was stable over time (eFigure 1 in Supplement 1). For individuals with autism claims, 71.6% were continuously enrolled in Medicaid (ie, no coverage gaps >6 months). In 2019, 46% of enrollees with autism claims had claims for intellectual disability, down from 62% in 2011 (eFigure 2 in Supplement 1).
Table. Characteristics of Adults With Medicaid Claims for Autism and Adults Without Medicaid Claims for Developmental Disability, 2011-2019.
| Autism, No. (%) (n = 403 028) | Full Medicaid denominator (n ≈ 37 million per y)a | |
|---|---|---|
| Sex | ||
| Female | 103 725 (25.7) | 56.2 |
| Male | 299 301 (74.3) | 43.8 |
| Race | ||
| Asian | 12 719 (3.3) | 6.1 |
| Black | 65.046 (16.8) | 23.7 |
| Native American | 2914 (0.8) | 1.3 |
| Pacific Islander | 3126 (0.8) | 2.2 |
| White | 287 779 (74.2) | 61.3 |
| Multiple races | 16.186 (4.2) | 5.4 |
| Missing | 15 258 | NA |
| Ethnicity | ||
| Non-Hispanic or Latino | 340 579 (87.8) | 79.3 |
| Hispanic or Latino | 47 191 (12.2) | 20.7 |
| Missing | 15 258 | NA |
| Region | ||
| Northeast | 111 179 (28.0) | 20.1 |
| Midwest | 94 627 (23.5) | 19.7 |
| South | 114 579 (28.4) | 32.2 |
| West | 81 411 (20.2) | 25.3 |
| Other or US territory | 1230 (0.3) | 1.4 |
| Eligibility type (ever)b | ||
| Disability | 236 778 (58.7) | 24.1 |
| Income | 254 019 (63.0) | 53.4 |
| Age at first claim, y | ||
| 18-24 | 280 911 (69.7) | 34.2 |
| 25-34 | 56 481 (14.0) | 18.5 |
| 35-44 | 28 264 (7.0) | 13.6 |
| 45-54 | 21 671 (5.4) | 12.0 |
| 55-64 | 11 643 (2.9) | 10.9 |
| 65-89 | 4058 (1.0) | 10.7 |
| Years enrolled per person | ||
| Mean (SD) | 5.8 (2.8) | 2.7 (2.6) |
| Median (range) | 6.0 (0.0-9.0) | 3.3 (0.0-9.0) |
| Year enrolled | ||
| 2011 | 149 861 (37.0) | 32.2 |
| 2012 | 168 837 (41.8) | 33.2 |
| 2013 | 190 168 (47.0) | 34.0 |
| 2014 | 228 664 (56.7) | 42.0 |
| 2015 | 264 497 (65.6) | 48.0 |
| 2016 | 295 826 (73.4) | 49.8 |
| 2017 | 320 241 (79.5) | 50.6 |
| 2018 | 343 948 (85.3) | 51.3 |
| 2019 | 360 019 (89.3) | 51.6 |
| Continuously enrolled | 288.533 (71.6) | 30.0 |
Abbreviation: NA, not applicable.
Full denominator calculated using the Centers for Medicare & Medicaid Services annual data book12 weighted by demographic distribution in our approximate 1% Medicaid sample excluding individuals with intellectual disability (about 1.5% of full Medicaid sample).
Eligibility type for Medicaid can change over time. Variable categorized as ever eligible based on disability and ever eligible based on income; thus, values do not sum to 1.
Total prevalence increased from 4.2 per 1000 enrollees in 2011 to 9.5 in 2019 (Figure 1). The largest increase was observed in the 25- to 34-year age group (195%), and the smallest was in the 55- to 64-year age group (45%). In quantitative bias analysis, an imperfect negative predictive value could lead to an increased underlying prevalence by more than 10 cases per 1000 across age groups (eFigure 3 in Supplement 1).
Figure 1. Autism Prevalence in Medicaid-Enrolled Adults Aged 18 Years or Older, 2011-2019.
Autism was most prevalent among White enrollees aged 18 to 25 years (20.5 per 1000 enrollees), and all groups showed decreasing prevalence with increasing age. Prevalence in non-Hispanic adults aged 18 to 25 years was 1.8 times that of Hispanic adults (18.7 vs 10.4 per 1000 enrollees) (Figure 2). The relative difference comparing prevalence in non-Hispanic with Hispanic enrollees increased with increasing age.
Figure 2. Autism Prevalence in Medicaid-Enrolled Adults by Race and Ethnicity Stratified by Age, 2019.
Multiple imputation models were used to account for missing race and ethnicity data. Multiple races indicate having 2 different races reported in any 2 years.
Discussion
This cohort study found that the prevalence of autism in adults, as represented by health care claims, rose in the Medicaid system from 2011 to 2019. Our results extend the work of Schott et al,6 who characterized prevalence from 2008 to 2012, finding continued increasing prevalence. The demographic characteristics of the adult Medicaid population mirror those of other epidemiologic studies of autism in this period, eg, majority are male, White, and not Hispanic vs the general population.1
By age, there was a sharp increase in prevalence in the youngest 2 age groups, which corresponds to studies in successive cohorts of children.1 With more than 70% continuous enrollment within this population, we anticipate corresponding increases in prevalence in older age groups as individuals age. Our results align with the decreasing co-occurrence of intellectual disability over time, highlighting improved diagnostic approaches and awareness of autism without intellectual disability.14
Our findings showed consistent differences in the identified prevalence of autism by racial group. This aligns with work highlighting the diagnosis gap in children, where Black and other children from minoritized groups were not less likely to be on the spectrum but were less likely to be identified.8 The racial gap is independent of income, but income does correlate with disparities in identification and race.8 While this pattern has diminished among children,2 it is still seen in adults identified under previous inequitable practices.
Limitations
This study is limited by reliance on ICD codes for autism. Our prevalence calculations estimate the number of adults identified with autism and served using Medicaid funds, but many undiagnosed adults in the older age ranges could exist.15 We have estimated the range of prevalence accounting for undiagnosed adults in our bias analysis. Many adults are enrolled for short periods and may not have enough claims to assess for autism; however, our quantitative bias analysis presents possible associations of such bias with prevalence.
Conclusions
This cohort study’s findings suggest that as children on the autism spectrum become autistic adults, Medicaid will become increasingly important as an insurance provider. Medicaid claims data may be a valuable resource for understanding the health of the autistic population in the future.
eTable. International Classification of Diseases, Ninth Revision and Tenth Revision Codes Used to Identify Autism
eMethods. Additional Data Processing Detail
eReferences
eFigure 1. Percentage of Autistic Adults by Sex, 2011-2019
eFigure 2. Percentage of Autistic Adults With Intellectual Disability, 2011-2019
eFigure 3. Quantitative Bias Analysis for Adult Autism Prevalence Estimates in Medicaid Claims From 2011 to 2019, Accounting for Imperfect Positive and Negative Predictive Values of Autism Claims by Age Category
Data Sharing Statement
References
- 1.Maenner MJ, Shaw KA, Bakian AV, et al. Prevalence and characteristics of autism spectrum disorder among children aged 8 years. MMWR Surveill Summ. 2021;70(11):1-16. doi: 10.15585/mmwr.ss7011a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Maenner MJ, Warren Z, Williams AR, et al. Prevalence and characteristics of autism spectrum disorder among children aged 8 years. MMWR Surveill Summ. 2023;72(2):1-14. doi: 10.15585/mmwr.ss7202a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mandell DS, Barry CL. Care for autism and other disabilities. N Engl J Med. 2017;376(10):e15. doi: 10.1056/NEJMp1700697 [DOI] [PubMed] [Google Scholar]
- 4.Gilmore D, Krantz M, Weaver L, Hand BN. Healthcare service use patterns among autistic adults. Autism. 2022;26(2):317-331. doi: 10.1177/13623613211060906 [DOI] [PubMed] [Google Scholar]
- 5.Rubenstein E, Bishop L. Is the autism boom headed for Medicaid? Autism Res. 2019;12(10):1541-1550. doi: 10.1002/aur.2173 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Schott W, Verstreate K, Tao S, Shea L. Autism grows up. Psychiatr Serv. 2021;72(5):597. doi: 10.1176/appi.ps.202000144 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Anderson KA, Hemmeter J, Rast JE, Roux AM, Shattuck PT. Trends in supplemental security income payments to adults with autism. Psychiatr Serv. 2020;71(6):602-607. doi: 10.1176/appi.ps.201900265 [DOI] [PubMed] [Google Scholar]
- 8.Mandell DS, Wiggins LD, Carpenter LA, et al. Racial/ethnic disparities in the identification of children with autism spectrum disorders. Am J Public Health. 2009;99(3):493-498. doi: 10.2105/AJPH.2007.131243 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Rubenstein E, Daniels J, Schieve LA, et al. Trends in special education eligibility among children with autism spectrum disorder, 2002-2010. Public Health Rep. 2018;133(1):85-92. doi: 10.1177/0033354917739582 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Benevides TW, Carretta HJ, Rust G, Shea L. Racial and ethnic disparities in benefits eligibility and spending among adults on the autism spectrum. PLoS One. 2021;16(5):e0251353. doi: 10.1371/journal.pone.0251353 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zhang G, Rose CE, Zhang Y, et al. Multiple imputation of missing race and ethnicity in CDC COVID-19 case-level surveillance data. Int J Stat Med Res. 2022;11:1-11. doi: 10.6000/1929-6029.2022.11.01 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Center for Medicare and Medicaid Services. Data that helps you better understand CMS programs. Accessed February 21, 2023. http://data.cms.gov
- 13.Lash TL, Fox MH, Fink AK. Applying Quantitative Bias Analysis to Epidemiologic Data. Springer; 2009. doi: 10.1007/978-0-387-87959-8 [DOI] [Google Scholar]
- 14.Coo H, Ouellette-Kuntz H, Lloyd JE, Kasmara L, Holden JJ, Lewis ME. Trends in autism prevalence: diagnostic substitution revisited. J Autism Dev Disord. 2008;38(6):1036-1046. doi: 10.1007/s10803-007-0478-x [DOI] [PubMed] [Google Scholar]
- 15.Lai MC, Baron-Cohen S. Identifying the lost generation of adults with autism spectrum conditions. Lancet Psychiatry. 2015;2(11):1013-1027. doi: 10.1016/S2215-0366(15)00277-1 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable. International Classification of Diseases, Ninth Revision and Tenth Revision Codes Used to Identify Autism
eMethods. Additional Data Processing Detail
eReferences
eFigure 1. Percentage of Autistic Adults by Sex, 2011-2019
eFigure 2. Percentage of Autistic Adults With Intellectual Disability, 2011-2019
eFigure 3. Quantitative Bias Analysis for Adult Autism Prevalence Estimates in Medicaid Claims From 2011 to 2019, Accounting for Imperfect Positive and Negative Predictive Values of Autism Claims by Age Category
Data Sharing Statement