Key PointsQuestion
What are the rates of comorbid mood and anxiety disorders in individuals with autism spectrum disorder?
Findings
In the cohort study of 31 220 individuals born in Olmsted County, Minnesota, those who were identified as having autism spectrum disorder were significantly more likely to have comorbid depression, anxiety, and bipolar disorder compared with age- and sex-matched referents.
Meaning
The findings suggest that individuals with autism spectrum disorder may be more likely to receive diagnoses of depression, bipolar disorder, and anxiety than those without an autism spectrum disorder diagnosis.
Importance
It is critical to evaluate the risk of comorbid psychiatric diagnoses to meet the needs of individuals with autism spectrum disorder (ASD).
Objective
To examine whether individuals with ASD are at greater risk for comorbid diagnoses of depression, anxiety, or bipolar disorder.
Design, Setting, and Participants
This cohort study used data from a population-based birth cohort of 31 220 individuals born in Olmsted County, Minnesota, from January 1, 1976, to December 31, 2000. Patients with research-identified ASD were previously identified using a multistep process that evaluated signs and symptoms abstracted from medical and educational records. For each of the 1014 patients with ASD, 2 age- and sex-matched referents who did not meet criteria for ASD were randomly selected from the birth cohort (n = 2028). Diagnosis codes for anxiety, depression, and bipolar disorders were electronically obtained using the Rochester Epidemiological Project records-linkage system. Data analysis was performed from July 1, 2018, to April 1, 2019.
Main Outcomes and Measures
Cumulative incidence of clinically diagnosed depression, anxiety, and bipolar disorder through early adulthood in individuals with ASD compared with referents.
Results
A total of 1014 patients with ASD (median age at last follow-up, 22.8 years [interquartile range, 18.4-28.0 years]; 747 [73.7%] male; 902 [89.0%] white) and 2028 referents (median age at last follow-up, 22.4 years [interquartile range, 18.8-26.2 years]; 1494 [73.7%] male; 1780 [87.8%] white) participated in the study. Patients with ASD were significantly more likely to have clinically diagnosed bipolar disorder (hazard ratio [HR], 9.34; 95% CI, 4.57-19.06), depression (HR, 2.81; 95% CI, 2.45-3.22), and anxiety (HR, 3.45; 95% CI, 2.96-4.01) compared with referents. Among individuals with ASD, the estimates of cumulative incidence by 30 years of age were 7.3% (95% CI, 4.8%-9.7%) for bipolar disorder, 54.1% (95% CI, 49.8%-58.0%) for depression, and 50.0% (95% CI, 46.0%-53.7%) for anxiety. Among referents, cumulative incidence estimates by 30 years of age were 0.9% (95% CI, 0.1%-1.7%) for bipolar disorder, 28.9% (95% CI, 25.7%-32.0%) for depression, and 22.2% (95% CI, 19.3%-25.0%) for anxiety.
Conclusions and Relevance
The findings suggest that individuals with ASD may be at increased risk for clinically diagnosed depression, anxiety, and bipolar disorder compared with age- and sex-matched referents. This study supports the importance of early, ongoing surveillance and targeted treatments to address the psychiatric needs of individuals with ASD.
Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by deficits in social communication skills and a pattern of restricted or repetitive interests and behaviors.1 Recent ASD prevalence estimates in the United States range from 1.68% among individuals aged 8 years in an active public health surveillance program to 2.41% in individuals aged 3 to 17 years based on a national parent survey, with male to female ratios of 4.0 in the group of 8-year-old individuals and 2.9 in the group aged 3 to 17 years.2 Increased risk of psychiatric comorbidity has been widely reported3,4; however, estimates of common comorbidities, such as depression, anxiety, or bipolar disorder, range widely.5-7 Diverse methods and varied, often small samples (eg, inpatient unit, clinical samples) have yielded discrepant findings about the risk of psychiatric comorbidity in individuals with ASD.8-10
A few studies have examined psychiatric comorbidity in individuals with ASD using large-scale research databases11 or larger epidemiologic studies to identify samples for further assessment.12,13 Even fewer studies have used population-based techniques. Specifically, using electronic medical records from 4 hospitals in a US city, Kohane and colleagues14 examined the comorbidity of various medical conditions (eg, autoimmune conditions, epilepsy, and Down syndrome) in individuals with ASD; however, their examination of psychiatric comorbidity was limited to schizophrenia.14 Schendel and colleagues15 examined comorbidity using a longitudinal cohort study of children born in Denmark. Although their primary analyses were focused on mortality rates, the demographic tables in their research highlighted an increased risk of psychiatric comorbidities in individuals with ASD; however, broad diagnostic categories (ie, mood [affective] disorders and neurotic, stress-related, somatoform disorders) did not yield clear findings about the risk for specific psychiatric concerns.15 In a Swedish population-based cohort study, Rai and colleagues16 assessed solely the risk of depression and found an increased risk among individuals with ASD in general, with greater risk among individuals with ASD without intellectual disability than among those with ASD and intellectual disability.
Because psychiatric symptoms have a known association with functioning in social, academic, and vocational environments, understanding the risk of comorbid mood and anxiety disorders is critical to improve long-term outcomes for individuals with ASD. Knowledge of the population-based risk of associated psychiatric conditions may facilitate advances in necessary screening and treatment guidelines. The current study used a large, longitudinal, population-based birth cohort to examine the risk of clinically diagnosed depression, anxiety, and bipolar disorders during childhood through early adulthood in individuals with ASD compared with age- and sex-matched referents from the same birth cohort.
The REP Medical Record Linkage System
The study population for this cohort study was assembled using birth certificate data available for Olmsted County, Minnesota, obtained from the Minnesota Department of Health and using the resources of the Rochester Epidemiology Project (REP). The REP is a medical records-linkage system that includes records of all outpatient and inpatient medical professionals in the community, including Mayo Clinic, Olmsted Medical Center, their 3 affiliated hospitals, and several smaller care practices.17-19 The REP provides centralized longitudinal medical data, including clinical documentation from primary care and specialty clinics, emergency department visits, hospitalizations, laboratory results, social services reports, and birth and death certificate data. Diagnoses assigned at each encounter are coded and maintained in continuously updated electronic files. The study was approved by the institutional review boards at Olmsted Medical Center and Mayo Clinic. Access to school records was made possible through contractual agreement of Mayo Clinic, the Independent School District No. 535 school board, and equivalent authorities governing private schools. We had a contractual agreement with the school district to review the records. No consent was obtained from the individuals, but consent was from the school board; however, if a patient had denied access to medical records, the patient was excluded entirely. The data were not deidentified because of the need to match the student records with the medical records.
Quiz Ref IDThe study population included 43 215 children born from January 1, 1976, to December 31, 2000, whose mothers were residents of Olmsted County, Minnesota, during the child’s birth. Of the 43 215 children, 39 890 had medical records available for research purposes in accordance with Minnesota state privacy law, statute 144.335. Of these, 31 220 were still living in Olmsted County at 3 years of age20,21 and considered for this study.
Identification of Patients With ASD
Quiz Ref IDThe ASD case status was determined using data from the REP and school documentation as described in detail previously.21 The first step was identifying relevant neurodevelopmental or psychiatric disorder (NPD) diagnostic codes from medical records and educational disability classification codes from school records that overlap the core social, communicative, and behavioral features of ASD or commonly coexist with ASD. These codes were grouped into NPD clusters (ie, ASD, childhood psychosis, developmental and cognitive, speech and language, nonpsychosis and childhood psychiatric, and neurologic and motor).21 In the second step, of the 4301 individuals with codes in NPD clusters, 1766 were identified as having potential ASD if they had diagnoses or educational disability classification codes in the ASD cluster, diagnoses or educational disability classification codes in 3 or more different NPD clusters, or codes in only the developmental and cognitive or speech and language cluster or in 2 NPD clusters and screened positive based on further review. In the third step, medical or school records of all individuals from the potential ASD group were manually reviewed to abstract data through a systematic, multistaged process, allowing accumulation of all needed information for each potential ASD case. Quiz Ref IDThe abstractors carefully read all documentation of each medical and school record through 21 years of age to identify descriptive phrases using a data dictionary developed by the research team that mapped to any of the 58 ASD signs or symptoms that contributed to the 12 Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR) criteria.22 This approach yielded 1296 individuals who had at least 2 social criteria and at least 1 communication or restricted repetitive and stereotyped patterns of behavior, interests, and activities criterion and were classified as having research-identified ASD. Patients with false-positive diagnoses were excluded by manual review to identify individuals whose signs or symptoms were related to other conditions, such as psychosis or major depression.21 Of the 1056 patients with research-identified ASD incident cases who remained after manual review, 1014 had not denied research authorization.
Identification of Age- and Sex-Matched Controls
For each of the 1014 patients with ASD, 2 age- and sex-matched referents who did not meet criteria for research-identified ASD were identified from those with research authorization in the larger birth cohort. The referents were randomly selected from the pool of individuals of the same sex who were born within 30 days of the patients with ASD and who were still in the community at the time when the patient with ASD met the research criteria for ASD.
Identification of Depression, Anxiety, and Bipolar Disorder
Among the 1014 patients with ASD and 2028 matched referents, clinical diagnostic codes for depression, anxiety, and bipolar disorder assigned by an REP-affiliated practitioner through June 30, 2017, were electronically obtained. Clinical diagnoses from January 1, 1976, through September 31, 2015, were originally coded using 2 different coding systems, Hospital Adaptation of the International Classification of Diseases, Eighth Revision (HICDA) and International Classification of Diseases, Ninth Revision (ICD-9), depending on diagnosis date and source. Beginning in October 2015, diagnoses were coded using International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes. Diagnosis codes were reviewed by 4 child psychologists or psychiatrists and categorized as (1) anxiety-related diagnoses, (2) depression-related diagnoses, or (3) bipolar disorder–related diagnoses. For a full list of diagnoses included in these categories, see the eTable in the Supplement. Individuals were classified by the study statistician (A.L.W.), blinded to the ASD status, as having a diagnosis of depression if they had 2 or more visit dates greater than 30 days apart with depression-related diagnosis codes. The same approach was used for classifying individuals as having anxiety and bipolar disorder. In addition, the subset of medical records for individuals who met any of the following criteria were manually reviewed by the practitioners to exclude those without reasonable evidence in their medical record of the specific diagnosis category classification: (1) first diagnosed with depression, anxiety, or bipolar disorder before 5 years of age; (2) had only 2 visit dates greater than 30 days apart with diagnosis codes from the same category; (3) had only 3 or 4 visit dates with bipolar disorder–related diagnosis codes; (4) met classification criteria for depression but also had a single visit with an anxiety-related diagnosis code or 2 or more visits that were not 30 days apart with anxiety-related diagnosis codes; or (5) did not meet classification criteria for depression but had suicide-related diagnosis codes. Individuals could be classified as having more than 1 comorbid diagnosis.
Data analysis was performed from July 1, 2018, to April 1, 2019. Analyses were performed using SAS software, version 9.4 (SAS Institute Inc). Separate analyses were performed for each of the 3 psychiatric comorbidities (ie, events) of interest (ie, depression, anxiety, and bipolar disorder). For individuals who were classified as having an event, their duration of follow-up was calculated from birth to the date of their first diagnosis of the psychiatric comorbidity; for individuals who did not meet the criteria for a diagnosis, their duration of follow-up was censored at the date of their last visit to an REP-affiliated practitioner before June 30, 2017. For each group (patients with ASD and matched referents), the event-free survival was estimated using the Kaplan-Meier method, which takes into account the varying duration of follow-up for each individual, converted to 100 minus the Kaplan-Meier estimate, and is herein referred to as the cumulative incidence of the event. This calculation did not take into account the competing risk of death, but only 0.8% of the patients with ASD and matched referents without any of the 3 psychiatric comorbidities were deceased at the time of this study. Cox proportional hazards regression models were fit to estimate the association between ASD status and meeting criteria for the comorbidity of interest; associations were summarized using hazard ratios (HRs) and corresponding 95% CIs.
As a sensitivity analysis, main analyses were also completed within a more narrowly defined group that used stricter inclusion criteria for ASD research diagnosis, as described previously.21 The narrow, more conservative research definition required that at least 6 criteria were met, including at least 2 social interaction criteria, 1 communication criterion, and 1 restricted, repetitive behavior criterion.
Demographics and Follow-up
A total of 1014 patients with ASD (median age at last follow-up, 22.8 years [interquartile range, 18.4-28.0 years]; 747 [73.7%] male; 902 [89.0%] white) and 2028 referents (median age at last follow-up, 22.4 years [interquartile range, 18.8-26.2 years]; 1494 [73.7%] male; 1780 [87.8%] white) participated in the study. Table 1 summarizes demographic characteristics at birth for the 1014 patients with ASD and 2028 age- and sex-matched referents.
Among individuals with ASD, estimates of cumulative incidence by 30 years of age were 7.3% (95% CI, 4.8%-9.7%) for bipolar disorder, 54.1% (95% CI, 49.8%-58.0%) for depression, and 50.0% (95% CI, 46.0%-53.7%) for anxiety. Among referents, the estimates of cumulative incidence by 30 years of age were 0.9% (95% CI, 0.1%-1.7%) for bipolar disorder, 28.9% (95% CI, 25.7%-32.0%) for depression, and 22.2% (95% CI, 19.3%-25.0%) for anxiety (Figure 1). Individuals with ASD were significantly more likely than referents to have comorbid clinically diagnosed bipolar disorder (HR, 9.34; 95% CI, 4.57-19.06), depression (HR, 2.81; 95% CI, 2.45-3.22), and anxiety (HR, 3.45; 95% CI, 2.96-4.01) (Table 2).
Quiz Ref ID Compared with the referents, individuals with ASD received clinical diagnoses at a younger age for depression (median [IQR] age, 18.1 [15.8-22.5] vs 15.7 [12.7-19.9] years; P < .001) and anxiety (median [IQR] age, 20.3 [15.8-24.2] vs 15.2 [10.9-20.5] years; P < .001). Bipolar disorder was also diagnosed at a younger age among individuals with ASD (median [IQR] age, 27.0 [17.2-28.3] vs 20.3 [16.2-25.7] years; P = .20); however, the statistical power was limited, with only 9 referents clinically diagnosed with bipolar disorder, and the finding was not statistically significant.
Analyses Stratified by Sex
When the results were stratified by sex, increased risk of each psychiatric comorbidity in individuals with ASD compared with matched referents was observed for both males and females (Table 2 and Figure 2). Although the increased risk of clinically diagnosed bipolar disorder in individuals with ASD compared with matched referents was higher among males (HR, 12.14; 95% CI, 4.74-31.10) compared with females (HR, 5.85; 95% CI, 1.92-17.84), the difference in risk was not statistically significant (test for interaction, P = .33) in part because of the small number with bipolar disorder diagnoses. The increased risk of clinically diagnosed depression (HR, 3.17 [95% CI, 2.67-3.77] vs 2.28 [95% CI, 1.82-2.86]; P = .01) and anxiety (HR, 3.87 [95% CI, 3.19-4.70] vs 2.91 [95% CI, 2.28-3.71]; P = .05) in individuals with ASD compared with matched referents was significantly higher among males than females despite females with and without ASD having greater cumulative incidence of comorbidities (described below).
As shown in Figure 2C, the cumulative incidence of clinically diagnosed anxiety was higher in females compared with males, both among individuals with ASD (HR, 1.49; 95% CI, 1.22-1.83) and among referents (HR, 1.99; 95% CI, 1.57-2.52). A similar finding was also observed for clinically diagnosed depression (patients with ASD: HR, 1.35 [95% CI, 1.11-1.64]; referents: HR, 1.96 [95% CI, 1.59-2.40]) (Figure 2B). However, among individuals with ASD, before 15 years of age, the cumulative incidence of clinically diagnosed depression was initially higher in males compared with females, but after 15 years of age, the cumulative incidence of depression was higher in females.
Multiple Psychiatric Comorbidities
Among the individuals who met criteria for at least 1 of the 3 psychiatric comorbidities, those with ASD were more likely than referents to meet criteria for 2 or more comorbidities (318 of 574 [55.3%] vs 202 of 457 [44.2%], P < .001). Quiz Ref IDOf the 573 individuals with ASD who met criteria for at least 1 of the 3 psychiatric comorbidities, 256 (44.7%) met criteria for only 1 diagnosis, 290 (50.6%) met criteria for 2 diagnoses, and 27 (4.7%) met criteria for all 3 at some point in time. In contrast, of the 457 referents who met criteria for at least 1 of the 3 psychiatric comorbidities, 255 (55.8%) met criteria for only 1 diagnosis, 196 (42.9%) for 2 diagnoses, and 6 (1.3%) for all 3 diagnoses.
Among the 1014 patients with ASD, 510 met criteria for the narrow, more conservative research definition. Greater risk of psychiatric comorbidities was observed when the analysis was restricted to patients meeting these stricter criteria (Table 3). Estimates of cumulative incidence by 30 years of age were 7.4% (95% CI, 3.6%-11.1%) for bipolar disorder, 53.7% (95% CI, 47.1%-59.6%) for depression, and 55.6% (95% CI, 49.3%-61.2%) for anxiety compared with 1.4% (95% CI, 0%-3.2%) for bipolar disorder, 30.9% (95% CI, 25.6%-35.8%) for depression, and 24.7% (95% CI, 19.9%-29.3%) for anxiety in the referent group (eFigure in the Supplement).
Although previous studies15,23 have used population techniques to examine medical or psychiatric risk more broadly, this is the first study, to our knowledge, to use a population-based birth cohort to longitudinally assess the risk of multiple psychiatric comorbidities, specifically clinically diagnosed depression, anxiety, and bipolar disorder, among individuals with ASD into middle adulthood. Current results suggest that individuals with ASD may be at increased risk for anxiety and mood diagnoses compared with those without ASD, with estimates falling within the wide range of estimated risk reported in prior studies8-10 that used different methods and samples. The current results suggest an even greater risk of clinically diagnosed depression among individuals with ASD in the US county sampled in this study compared with recent research conducted in Sweden, which found that 19.8% of individuals with ASD were diagnosed with depression by 27 years of age.16
Significant psychosocial sequelae associated with having ASD, including difficulties developing and maintaining relationships, challenges succeeding academically and vocationally, and behaviors that can be problematic to manage, particularly increase risk for mood and anxiety symptoms in individuals with ASD.24-27 Individuals with ASD also experience greater rates of other mental health challenges, including attention-deficit/hyperactivity disorder and substance abuse.13,28 These comorbidities, as well unique factors associated with autism, are believed to contribute to greater suicide risk in the population with ASD.29 In addition, potential shared etiologic factors among mood disorders, anxiety disorders, and ASD may contribute to high levels of comorbidity.
Sex appears to play an important role in the association between ASD and depression and anxiety. Specifically, although having ASD was associated with an increased rate of depression and anxiety among females, females with and without ASD experienced higher rates of these diagnoses compared with their male counterparts. Being female with ASD may be associated with a greater risk for psychiatric problems compared with being female in general. However, the association between having ASD and comorbid depression and anxiety was stronger among males than females despite males having lower rates of comorbidity overall, which is especially salient because a greater proportion of individuals with ASD are male.30
In addition, individuals with ASD were diagnosed earlier with depression, anxiety, and bipolar disorder, and individuals with ASD were more likely to receive multiple psychiatric diagnoses than individuals without ASD. Of importance, individuals with ASD may present with a different trajectory of symptoms that necessitate ongoing screening starting at younger ages. This finding may also reflect that individuals with ASD may be monitored more closely medically and more readily connected to resources appropriate for screening and diagnosis given their ongoing need for services. Although individuals with ASD may receive these diagnoses at younger ages, they may continue to have greater risk in adulthood, when access to specialty care and resources is more limited, which may place a large burden for screening on primary care practitioners providing care to adults with ASD.31
The core features of ASD, including language challenges, problems labeling and describing emotions, and difficulty engaging socially in a reciprocal manner, may make psychiatric diagnosis and assessment of response to treatment particularly challenging and necessitate development of novel diagnostic tools and outcome measures.32 Given the high rates of comorbidity, researchers and practitioners should develop tools that are specific to the unique needs of this population and effective medications and treatments for mood and anxiety concerns, which remain limited in this population.33
This study has limitations. Overall, the population of Olmsted County, Minnesota, is generally more educated, wealthier, and less racially diverse than the broader United States, potentially limiting generalizability.19 In addition, individuals were identified as having ASD based on research criteria using DSM-IV-TR criteria and retrospective record review without direct assessment to validate the diagnosis.21 Diagnoses of anxiety, depression, and bipolar disorder were discerned from medical records, which may result in overrepresentation and underrepresentation; however, that, as well as changes in diagnostic practices, should be equal across patients with ASD and age- and sex-matched referents. The broad nature of the overarching diagnostic categories of interest does not provide information about the risk of specific diagnoses, which will be an important avenue for future research.
The findings suggest that individuals with ASD may be at increased risk for clinically diagnosed depression, anxiety, and bipolar disorder compared with age- and sex-matched referents. These findings support early, ongoing surveillance for psychiatric comorbidities in the population with ASD because identifying comorbidities may guide intervention and improve quality of life for individuals with ASD.
Accepted for Publication: July 19, 2019.
Corresponding Author: Alexandra C. Kirsch, PhD, Department of Psychiatry and Psychology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905 (kirsch.alexandra@mayo.edu).
Published Online: December 2, 2019. doi:10.1001/jamapediatrics.2019.4368
Author Contributions: Dr Kirsch and Ms Weaver 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.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: Kirsch, Huebner, Mehta, Weaver, Myers, Voight, Katusic.
Drafting of the manuscript: Kirsch, Huebner, Weaver.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Howie, Weaver.
Obtained funding: Myers, Katusic.
Administrative, technical, or material support: Huebner, Mehta, Voight.
Supervision: Kirsch, Huebner, Mehta, Voight, Katusic.
Conflict of Interest Disclosures: Dr Kirsch reported receiving grants from the National Institutes of Health and the Public Health Service during the conduct of the study. Dr Huebner reported receiving grants from the National Institutes of Health during the conduct of the study. Dr Howie reported receiving grants from the National Institute of Mental Health during the conduct of the study. Dr Katusic reported receiving grants from the National Institutes of Health during the conduct of the study. No other disclosures were reported.
Funding/Support: This study was funded by research grants MH093522 from the National Institutes of Health and AG034676 from the US Public Health Service (Dr Katusic).
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
1.American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington, VA: American Psychiatric Publishing; 2013.
8.Albores-Gallo
L, Fritsche-García
L, Miranda-Aguirre
AP, Avila-Acosta
M. Brief report: macrocephaly phenotype and psychiatric comorbidity in a clinical sample of Mexican children and adolescents with autism spectrum disorders.
J Autism Dev Disord. 2017;47(9):2911-2917. doi:
10.1007/s10803-017-3175-4PubMedGoogle ScholarCrossref 9.Brookman-Frazee
L, Stadnick
N, Chlebowski
C, Baker-Ericzén
M, Ganger
W. Characterizing psychiatric comorbidity in children with autism spectrum disorder receiving publicly funded mental health services.
Autism. 2018;22(8):938-952. doi:
10.1177/1362361317712650PubMedGoogle ScholarCrossref 10.Joshi
G, Petty
C, Wozniak
J,
et al. The heavy burden of psychiatric comorbidity in youth with autism spectrum disorders: a large comparative study of a psychiatrically referred population.
J Autism Dev Disord. 2010;40(11):1361-1370. doi:
10.1007/s10803-010-0996-9PubMedGoogle ScholarCrossref 11.Rosenberg
RE, Kaufmann
WE, Law
JK, Law
PA. Parent report of community psychiatric comorbid diagnoses in autism spectrum disorders.
Autism Res Treat. 2011;2011:405849. doi:
10.1155/2011/405849PubMedGoogle Scholar 13.Simonoff
E, Pickles
A, Charman
T, Chandler
S, Loucas
T, Baird
G. Psychiatric disorders in children with autism spectrum disorders: prevalence, comorbidity, and associated factors in a population-derived sample.
J Am Acad Child Adolesc Psychiatry. 2008;47(8):921-929. doi:
10.1097/CHI.0b013e318179964fPubMedGoogle ScholarCrossref 19.St Sauver
JL, Grossardt
BR, Leibson
CL, Yawn
BP, Melton
LJ
III, Rocca
WA. Generalizability of epidemiological findings and public health decisions: an illustration from the Rochester Epidemiology Project.
Mayo Clin Proc. 2012;87(2):151-160. doi:
10.1016/j.mayocp.2011.11.009PubMedGoogle ScholarCrossref 20.Katusic
SK, Colligan
RC, Myers
SM,
et al. What can large population-based birth cohort study ask about past, present and future of children with disorders of development, learning and behaviour?
J Epidemiol Community Health. 2017;71(4):410-416. doi:
10.1136/jech-2016-208482PubMedGoogle ScholarCrossref 22.American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision. Arlington, VA: American Psychiatric Publishing; 2000.
25.Cheak-Zamora
NC, Teti
M, First
J. ‘Transitions are scary for our kids, and they’re scary for us’: family member and youth perspectives on the challenges of transitioning to adulthood with autism.
J Appl Res Intellect Disabil. 2015;28(6):548-560. doi:
10.1111/jar.12150PubMedGoogle ScholarCrossref 26.Hendricks
D. Employment and adults with autism spectrum disorders: challenges and strategies for success.
J Vocat Rehabil. 2010;32(2):125-134.
Google ScholarCrossref 30.Baio
J, Wiggins
L, Christensen
DL,
et al. Prevalence of autism spectrum disorder among children aged 8 years—autism and developmental disabilities monitoring network, 11 sites, United States, 2014.
MMWR Surveill Summ. 2018;67(6):1-23. doi:
10.15585/mmwr.ss6706a1PubMedGoogle ScholarCrossref 31.Vogan
V, Lake
JK, Tint
A, Weiss
JA, Lunsky
Y. Tracking health care service use and the experiences of adults with autism spectrum disorder without intellectual disability: a longitudinal study of service rates, barriers and satisfaction.
Disabil Health J. 2017;10(2):264-270. doi:
10.1016/j.dhjo.2016.11.002PubMedGoogle ScholarCrossref