CDC Report 57 Percent Increase in Autism

8/14/2019 CDC Report 57 Percent Increase in Autism

1/28

Morbidity and Mortality Weekly Reportwww.cdc.gov/mmwr

Surveillance Summaries December 18, 2009 / Vol. 58 / No. SS-10

Prevalence Auism SpecrumDisrders Auism and

Develpmenal DisabiliiesMniring Newrk,

Unied Saes, 2006

Department Of Health And Human Services

Centers for Disease Control and Prevention


2/28

MMWR

Te MMWR series o publications is published by Surveillance,Epidemiology, and Laboratory Services, Centers or Disease Controland Prevention (CDC), U.S. Department o Health and HumanServices, Atlanta, GA 30333.

Suggesed Ciain: Centers or Disease Control and Prevention.[itle]. Surveillance Summaries, [Date]. MMWR 2009;58(No. SS-#).

Ceners r Disease Cnrl and PreveninTomas R. Frieden, MD, MPH

Director

Peter A. Briss, MD, MPHActing Associate Director or Science

James W. Stephens, PhDOfce o the Associate Director or Science

Stephen B. Tacker, MD, MScActing Deputy Director or

Surveillance, Epidemiology, and Laboratory Services

Edirial and Prducin SaFrederic E. Shaw, MD, JD

Editor, MMWRSeries

Christine G. Casey, MDDeputy Editor, MMWRSeries

Susan F. Davis, MDAssociate Editor, MMWRSeries

eresa F. RutledgeManaging Editor, MMWRSeries

David C. JohnsonLead Technical Writer-Editor

Jerey D. Sokolow, MAProject Editor

Martha F. BoydLead Visual Inormation Specialist

Malbea A. LaPeteStephen R. Spriggserraye M. Starr

Visual Inormation Specialists

Kim L. BrightQuang M. Doan, MBA

Phyllis H. KingInormation Technology Specialists

Edirial BardWilliam L. Roper, MD, MPH, Chapel Hill, NC, Chairman

Virginia A. Caine, MD, Indianapolis, INJonathan E. Fielding, MD, MPH, MBA, Los Angeles, CA

David W. Fleming, MD, Seattle, WAWilliam E. Halperin, MD, DrPH, MPH, Newark, NJ

King K. Holmes, MD, PhD, Seattle, WADeborah Holtzman, PhD, Atlanta, GA

John K. Iglehart, Bethesda, MD

Dennis G. Maki, MD, Madison, WISue Mallonee, MPH, Oklahoma City, OK

Patricia Quinlisk, MD, MPH, Des Moines, IAPatrick L. Remington, MD, MPH, Madison, WI

Barbara K. Rimer, DrPH, Chapel Hill, NCJohn V. Rullan, MD, MPH, San Juan, PR

William Schaner, MD, Nashville, NAnne Schuchat, MD, Atlanta, GA

Dixie E. Snider, MD, MPH, Atlanta, GAJohn W. Ward, MD, Atlanta, GA

CoNtENtS

Introduction .............................................................................. 2

Methods ................................................................................... 3

Surveillance Methods .............................................................. 3

Study Sites ............................................................................. 3

Analytic Methods.................................................................... 6Results ...................................................................................... 7

Overall ASD Prevalence Estimates ........................................... 7

Prevalence by Sex and Race or Ethnicity................................... 8

Special Education Eligibility ..................................................... 9

Previously Documented Classifcation o ASD ............................ 9

Cognitive Functioning ............................................................. 9

Developmental Characteristics ................................................. 9

Comparison Between 2002 and 2006 Prevalence Estimates .... 10

Discussion............................................................................... 14

Changes in ASD Prevalence During 20022006..................... 15Factors Aecting Changes in Identifed ASD Prevalence .......... 15

Strengths o the ADDM Network Methodology ........................ 17

Limitations............................................................................ 17

Public Health Implications ...................................................... 17

Reerences .............................................................................. 19

Appendix ............................................................................... 21


3/28

Vol. 58 / SS-10 Surveillance Summaries 1

Prevalence of Autism Spectrum Disorders Autism and Developmental Disabilities Monitoring Network,

United States, 2006Autism and Developmental Disabilities Monitoring Network 2006 Principal Investigators

Abstract

Problem/Condition:Autism spectrum disorders (ASDs) are a group o developmental disabilities characterized by atypicaldevelopment in socialization, communication, and behavior. ASDs typically are apparent beore age 3 years, with associatedimpairments aecting multiple areas o a persons lie. Because no biologic marker exists or ASDs, identifcation is madeby proessionals who evaluate a childs developmental progress to identiy the presence o developmental disorders.

Reporting Period: 2006.

Methods: Earlier surveillance eorts indicated that age 8 years is a reasonable index age at which to monitor peak preva-lence. Te identifed prevalence o ASDs in U.S. children aged 8 years was estimated through a systematic retrospectivereview o evaluation records in multiple sites participating in the Autism and Developmental Disabilities Monitoring(ADDM) Network. Data were collected rom existing records in 11 ADDM Network sites (areas o Alabama, Arizona,

Colorado, Florida, Georgia, Maryland, Missouri, North Carolina, Pennsylvania, South Carolina, and Wisconsin)or2006. o analyze changes in identifed ASD prevalence, CDC compared the 2006 data with data collected rom 10sites (all sites noted above except Florida) in 2002. Children aged 8 years with a notation o an ASD or descriptionsconsistent with an ASD were identifed through screening and abstraction o existing health and education recordscontaining proessional assessments o the childs developmental progress at health-care or education acilities. Childrenaged 8 years whose parent(s) or legal guardian(s) resided in the respective areas in 2006 met the case defnition or anASD i their records documented behaviors consistent with the Diagnostic and Statistical Manual of Mental Disorders,4th edition, text revision (DSM-IV-R) criteria or autistic disorder, pervasive developmental disordernot otherwisespecifed (PDD NOS), or Asperger disorder. Presence o an identifed ASD was determined through a review o dataabstracted rom developmental evaluation records by trained clinician reviewers.

Results: For the 2006 surveillance year, 2,757 (0.9%) o 307,790 children aged 8 years residing in the 11 ADDM siteswere identifed as having an ASD, indicating an overall average prevalence o 9.0 per 1,000 population (95% confdence

interval [CI] = 8.69.3). ASD prevalence per 1,000 children aged 8 years ranged rom 4.2 in Florida to 12.1 in Arizonaand Missouri, with prevalence or the majority o sites ranging between 7.6 and 10.4. For 2006, ASD prevalence wassignifcantly lower in Florida (p


4/28

2 MMWR December 18, 2009

Public Health Actions: Tese results indicate an increased prevalence o identied ASDs among U.S. children aged 8years and underscore the need to regard ASDs as an urgent public health concern. Continued monitoring is needed todocument and understand changes over time, including the multiple ascertainment and potential risk actors likely tobe contributing. Research is needed to ascertain the actors that put certain persons at risk, and concerted eorts areessential to provide support or persons with ASDs, their amilies, and communities to improve long-term outcome.

InrducinAutism spectrum disorders (ASDs) are a group o develop-

mental disabilities characterized by atypical development insocialization, communication, and behavior. Te symptomso ASDs typically are present beore age 3 years and oten areaccompanied by abnormalities in cognitive unctioning, learn-ing, attention, and sensory processing (1). Te term spectrumdisorders is used to indicate that ASDs encompass a range obehaviorally dened conditions, which are diagnosed throughclinical observation o development. Tese conditions includeautistic disorder (i.e., autism), Asperger disorder, and pervasive

developmental disordernot otherwise specied (PDD-NOS)(24). Persons with Asperger disorder or PDD-NOS have ewerdiagnostic symptoms o ASDs compared with autism, and thesymptoms oten are indicative o more mild impairment. Tecomplex nature o these disorders, the current lack o consis-tent and reliable genetic or biologic diagnostic markers, andchanges in how these conditions are dened and identiedmake evaluating ASD prevalence over time challenging.

Since the early 1990s, the number o persons receiving ser-vices or ASDs has increased substantially (511). However,identiying children or services or autism might not beequivalent to using consistent diagnostic standards to identiy

persons in the population because services within communitiesare not available uniormly to all persons with ASDs. For thisreason, studies that rely exclusively on single-source administra-tive datasets (e.g., disability service records or annual reportso special education counts) most likely underestimate ASDprevalence and might not adequately capture changes in theASD population over time (8,1214).

Beore the 1980s, the term autism was used primarily toreer to autistic disorder and was thought to be rare, aect-ing approximately one in every 2,000 (0.5%) children (2,3).Autism now is considered to be one o three disorders classi-ed together as ASDs (4). Using diagnostic criteria establishedin the early 1990s, which encompass a broad spectrum odisorders (15,16), the best estimate o ASD prevalence is thatapproximately six or seven o every 1,000 (0.6%0.7%) chil-dren have an ASD. Tese estimates are approximately 10 timeshigher than estimates using earlier criteria (24,12). However,some recent population-based studies have documented evenhigher ASD prevalence estimates o >1% o children in areas

o Japan, Sweden, the United Kingdom, and the United State(12,1721), with ASD symptoms identied in 2.7% o chil-dren in one study rom Norway (22).

Elevated public concern, continued increases in the numbero children receiving services or ASDs, and reports o higher-than-expected ASD prevalence highlight the need or systematic public health monitoring o ASDs (23). For this reasonin 2000, CDC organized the Autism and DevelopmentalDisabilities Monitoring (ADDM) Network to provide a betteunderstanding o the prevalence, population characteristicsand public health impact o ASDs and other developmentaldisabilities in the United States (4,23). Te ADDM Networkemploys a multisite, multiple-source, records-based surveillancemethodology to conduct detailed retrospective screening andreview o behavioral data rom multiple education and healthacilities. Administrative records o children who have beenevaluated or a range o developmental conditions are reviewedand standard criteria or case identication and conrmationare applied at each o the surveillance sites (1,24).

wo surveillance summaries presenting ADDM NetworkASD prevalence data or 2000 and 2002 have been publishedpreviously (4,12). Tese rst two ADDM prevalence reportsserved to establish baseline prevalence or ASDs among U.S

children aged 8 years. Prevalence or those reports was calcu-lated by race/ethnicity, sex, and multiple ASD-associated char-acteristics (e.g., cognitive impairment) on the basis o data romsix sites* in 2000 and rom 14 sites in 2002. Overall prevalenceestimates or both surveillance years were comparable (6.7 and6.6 cases per 1,000 children aged 8 years in 2000 and 2002respectively). Among the participating sites, ASD prevalencein 2000 ranged rom 4.5 cases per 1,000 children aged 8 yearsin West Virginia to 9.9 cases in New Jersey, compared with 3.3cases in Alabama to 10.6 cases in New Jersey in 2002. From2000 and 2002, prevalence o identied cases o ASDs wasstable in our o the six sites (Arizona, Maryland, New Jersey

and South Carolina) that collected data in both years andincreased in two sites (Georgia and West Virginia).

* Areas o Arizona, Georgia, Maryland, New Jersey, South Carolina, and WesVirginia.

Areas o Alabama, Arizona, Arkansas, Colorado, Georgia, Maryland, MissouriNew Jersey, North Carolina, Pennsylvania, South Carolina, Utah, West Virginiaand Wisconsin.


5/28


Tis report presents results or 2006 rom 11 ADDM Networksites and describes updated ASD prevalence overall and byrace or ethnicity, sex, level o cognitive unctioning, ASD sub-type, and multiple associated characteristics. Because 10 sites(Alabama, Arizona, Colorado, Georgia, Maryland, Missouri,North Carolina, Pennsylvania, South Carolina, and Wisconsin)

also collected data on ASD prevalence in 2002 (12), changesin identied ASD prevalence in these sites are also reported. Inaddition, data or an additional surveillance year, 2004, werecollected by eight sites. Te 2004 surveillance year representedan eort on a smaller scale that was conducted by sites with avail-able resources to collect an additional surveillance year o data.Results or 2004 have been summarized (see Appendix).

Mehds

Surveillance MehdsTe ADDM Networks methodology (Figure 1) is modeled

on that used by CDCs Metropolitan Atlanta DevelopmentalDisabilities Surveillance Program (MADDSP), an active,population-based surveillance program that monitors theoccurrence o developmental disabilities through retrospectiverecord review among children aged 8 years in the metropolitanAtlanta area (1,2527). Although ASD symptoms typicallyare present in the rst 3 years o lie, identication oten isnot made not until later; previous surveillance eorts haveidentied age 8 years as a reasonable index age at which tomonitor peak prevalence or ASDs (1). Te ADDM Networkimplemented the MADDSP methodology using a commoncase denition and standardized data abstraction, clinicianreview, and quality assurance procedures.

Sudy SiesADDM site project teams are located at state health depart-

ments or at universities working on behal o their state healthdepartments to collect or receive inormation used or protect-ing public health. Sites were selected through a competitiveobjective review process on the basis o their ability to conductactive ASD records-based surveillance; they were not selected

to be a nationally representative sample. Each site met appli-cable local Institutional Review Board and/or other privacyand condentiality requirements.

Sudy Sies and Ppulain Characerisics

During 2006, CDC and 10 site project teams collaborated

in monitoring reported occurrence o ASDs in selected areaso 11 states (able 1). On the basis o postcensal estimatesthe number o children aged 8 years in the 11 surveillancesites ranged rom 7,184 in Colorado to 46,621 in Georgia(28). Distribution according to race/ethnicity among childrenaged 8 years varied across surveillance sites. Te percentageo non-Hispanic white children residing in each surveillancearea ranged rom 23.3% in Maryland to 69.7% in Alabama;the percentage o non-Hispanic black children ranged rom5.4% in Arizona to 50.1% in Pennsylvania; the percentage oHispanic children ranged rom 2.9% in Missouri to 52.3%in Florida; the percentage o Asian/ Pacic Islander children

ranged rom 1.1% in Alabama to 5.8% in Georgia; and the percentage o American Indian/Alaska native children was 0.6%in all sites except Arizona (with 2.2%). Te breakdown by sexwas similar across sites, with approximately equal distributiono male and emale children.

Case Defniin

For surveillance purposes, children who were born in 1998whose parent(s) or legal guardian(s) resided in the sites geographic region (able 1) at any time in 2006 were classiedas having an ASD i they displayed behaviors documentedin evaluation records by a community proessional that were

consistent with the Diagnostic and Statistical Manual o MentaDisorders, 4th Edition, Text Revision (DSM-IV-R) criteria (15)or the subtypes o autistic disorder, PDD-NOS (includingatypical autism), or Asperger disorder at any time through age8 years. An evaluation record was dened as the documentedrecord o an assessment conducted by a community proes-sional to determine the need or special education services orthe presence o a developmental disorder. Te assessmentscould be conducted at any time in the childs lie throughage 8 years. A community proessional was dened as a clini-cal or education proessional with specialized training in the

observation o children with developmental disabilities (e.g.a developmental pediatrician, child psychiatrist, pediatricneurologist, clinical or developmental psychologist, or speechor language pathologist). Te case denition ocuses on iden-tiying the behaviors consistent with the presence o any ASDrather than on attempting to identiy specic ASD subtypesChildren were classied as having cognitive impairment ithey had an intellectual quotient (IQ) score o 70 on theirmost recent test o intellectual ability available in the record

Areas o Alabama (32 counties in north and central Alabama), Arizona (onecounty, including metropolitan Phoenix), Colorado (1 county in metropolitanDenver), Florida (one southern county), Georgia (the CDC site in ive coun-ties in metropolitan Atlanta), Maryland (six counties in suburban Baltimore),Missouri (ive counties in metropolitan St. Louis), North Carolina (10 centralcounties), Pennsylvania (one county in metropolitan Philadelphia), SouthCarolina (23 counties in the Coastal and PeeDee regions), and Wisconsin (10counties in south eastern Wisconsin).

Areas o Alabama, Arizona, Georgia, Maryland, Missouri, North Carolina,South Carolina, and Wisconsin.


6/28


FIGURE 1. Surveillance methodology owchart Autism and Developmental Disabilities Monitoring (ADDM) Network, United States

Each site defines surveillance area by geographic boundaries and obtains data on the number andcharacteristics of resident children aged 8 years using the most recent vintage of postcensal population

estimates from CDC's National Center for Health Statistics.

Identify multiple health and education sources in the community that evaluate, educate,and treat children with developmental disabilities.

Obtain agreements for record review at these sources.

Request and receive data from health sourcesaccording to select ICD-9* and DSM-IV billing

codes.

Request and receive data from educationsources for all special education eligibility

classifications.

Import all data into database linking children'srecords across multiple sources to a unique

identifier per child.

Review and abstract individual health andeducation records in field and enter data

directly into database.

Replicate database at each site weekly andmerge abstracted data from multiple sources fora given child into one record; run reports for data

cleaning after each replication.

Complete abstraction for each child; recordsreviewed by trained, reliable clinician reviewers

to assign final case status.

Analyze data, and generate and disseminate reports to datasources, stakeholders, and scientific community for feedback

and distribution of information for public health action

Conduct ongoingquality control on a

10% sample of pendingcases for criticalclinician review fields.

Conduct ongoing qualitycontrol on the decision to

abstract for a 10%sample of records thatwere reviewed but not

abstracted and on a 10%sample of all abstractedrecords for critical data

fields; take both samplesfor each abstractor at

each site.

Implement final data cleaning procedures.

Link to vital statistics birth certificate data.

Submit data to ADDM pooled datasets.

* International Classifcation o Diseases, 9th Revision.Diagnostic and Statistical Manual o Mental Disorders, 4th Edition, Text Revision. All ADDM sites conducting cerebral palsy surveillance are conducting linkage o cases with vital statistics death certifcates. I easible, sites conductin

ASD and intellectual development surveillance also conducted this death certifcate linkage. For sites that completed this linkage no ASD cases weridentifed.

Georgia andNorth Carolina did not review special education records o children whose only exceptionality was speech and language impairment (SLI)A sample o children in SLI indicated that this decision had minimal eect on prevalence estimates.


7/28


8/28


or additional ICD-9 codes or other developmental disabili-ties monitored by that site (e.g., cerebral palsy or intellectualdisability). In addition, six sites (Arizona, Colorado, Georgia,Maryland, North Carolina, and South Carolina) (able 1)also included education sources and the special educationevaluation records o children receiving special education ser-

vices during 20052006 and/or 20062007. Each child wasassigned a unique identier to prevent duplication and to linkinormation across multiple data sources.

Investigators then reviewed the records o each provider oreducation source; records were selected or abstraction i thechild met the residency requirement and the record containeda documented or suspected ASD classication (i.e., a diagno-sis o an ASD or a special education eligibility classicationo autism) and/or descriptions o social behaviors associatedwith an ASD diagnosis. Abstracted inormation included theevaluation date; data source (education or health); communityproessional type and degree (e.g., MD, developmental pedia-trician or PhD, clinical psychologist); developmental historyand indications o developmental concerns (e.g., notationsin the records about parental or proessional worries that thechilds development was not progressing typically); verbatimbehavioral descriptions associated with autism (e.g., currentand past social, communication, and behavioral unctioning);results rom IQ, adaptive, and tests used to diagnose autism;and evaluation conclusions. Te most recent eligibility clas-sication or receiving special education services (e.g., autismor learning disability) was collected rom special educationrecords. For all abstracted evaluations, inormation rom

multiple sources was combined into one composite summaryrecord with a unique identier or each child and was thensent or clinician review.

Clinician Review

All abstracted evaluations rom the case ascertainmentphase were reviewed and scored by an ASD clinician reviewer.Clinician reviewers included proessionals(e.g., medical doc-tors, psychologists, and speech and language pathologists)with specialized training and experience in autism assessmentand diagnosis. Clinician reviewers were selected and assessedperiodically or reliability. Clinician reviewers used a coding

guide developed on the basis o DSM-IV-R criteria (15) todetermine whether each identied child met the ASD casedenition. A child met the ASD case denition i evidencedocumented in the abstracted evaluations indicated the pres-ence o autistic disorder, PDD-NOS, or Asperger disorder.Te clinician reviewer classied the child as having an ASD ievidence existed o either 1) the DSM-IV-Rcriteria in thesocial, communication, and behavior domains and evidence odelays beore age 3 years or 2) the social and either communica-

tion or behavior criteria or PDD-NOS or Asperger disorderBecause the thresholds or meeting the criteria specied byDSM-IV-R are lower or the diagnosis o PDD-NOS orAsperger disorder than or autistic disorder, a stricter require-ment was included requiring that at least one o the autism-specic behaviors be o a sucient quality or intensity to be

highly indicative o an ASD (12,23). Under this additionarequirement, or example, the DSM-IV-R social criteriono limited social or emotional reciprocity required a specicimpairment (e.g., rarely responds verbally or nonverbally toa social approach rom others in a amiliar setting).

In addition to the DSM-IV-R criteria, all evaluationinormation was reviewed to categorize any concerns noted indevelopmental evaluations concerning the childs developmentastatus beore age 3 years; any specied concerns regarding thedevelopment o social, language, or imaginative play beoreage 3 years also were documented, as were any indicationso regression or plateau in skill development. Descriptions oassociated eatures (e.g., odd responses to sensory stimuli) bythe community proessional also were coded. Te diagnosticconclusion o the community proessional who evaluated thechild also was summarized, including specication o any ASDdiagnosis by subtype, when available. Children were classied ashaving a previously documented ASD classication i they had1) received a diagnosis o autistic disorder, PDD-NOS, Aspergesyndrome, PDD, or ASD by a qualied proessional that wasdocumented in an evaluation record or 2) had received specialeducation services under an autism eligibility category.

Beore clinician review began or each surveillance year

interrater reliability was established among reviewers accord-ing to standards o 80% agreement or overall case status andother scored items. Ongoing inter-rater reliability checks wereconducted on a blinded, random sample o 10% o recordsreviewed. Te percentage o agreement or the nal case deni-tion was acceptable or 2006 (range: 82%100% across sites[Kappa range: 0.51.0]) and 2002 (12).

Analyic Mehds ASD prevalence estimates were calculated using as th

denominator the number o children aged 8 years residingin the study area according to CDCs National Center orHealth Statistics (NCHS) vintage 2007 postcensal populationestimates or each site (28). NCHS datasets provide estimatedpopulation counts by county, single year o age, race, ethnic ori-gin, and sex. Te race or ethnicity o each child was determinedrom inormation contained in the source records or, i notound in the source le, rom birth certicates (when available)Race- or ethnicity-specic rates were calculated or ve popu-lations: non-Hispanic white, non-Hispanic black, Hispanic


9/28


Asian/Pacic Islander, and American Indian/Alaska Native.Prevalence results are reported per 1,000 children aged 8 yearsidentied as meeting the ASD case denition over the totalnumber o children that age in the population. o allow com-parison o ASD estimates identied by community proessionalswith those identied by ADDM Network data, ASD prevalence

also was calculated on the basis o the number o children whohad a previously documented classication o an ASD on recordcompared with the total number o children in the populationaged 8 years. Poisson approximation to the binomial distribu-tion was used to calculate 95% condence intervals (CIs) orprevalence rates (30). Chi-square tests were used to compareprevalence estimates within and across sites, and rate ratios andpercentage change were used to compare prevalence changeswithin each site or 2002 and 2006 (31,32). Data rom 2002were compared with data rom 2006 because 2002 representedthe largest and most complete rst year o data collection or10 o the sites reporting in 2006. A maximum value o p


10/28


11/28


prevalence was signicantly higher among non-Hispanic whitechildren than among non-Hispanic black children (p


12/28


FIGURE 2. Overall prevalence* o autism spectrum disorders(ASDs) among children aged 8 years and prevalence o ASDsamong children with a previously documented ASD classifca-tion, by source type and order o ASD prevalence Autismand Developmental Disabilities Monitoring (ADDM) Network,11 sites, United States, 2006

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

ADDM 2006 ASD Prevalence

Prevalence

Florid

a

Alabam

a

Color

ado

Wisconsin

Penns

ylvania

South C

arolin

a

Maryl

and

Georgi

a

North C

arolin

a

Miss

ouri

Ariz

ona

Previously Documented ASD Classification Prevalence

0

* Per 1,000 population. Children were classifed as having a previously documented ASD clas-sifcation i they had 1) received a diagnosis o autistic disorder, pervasive

developmental disorder (PDD)not otherwise specifed, Asperger syn-drome, PDD, or ASD by a qualifed proessional that was documented inan evaluation record or 2) had received special education services underan autism eligibility category.

developmental plateau (i.e., lack o continued developmentwithout clear evidence o regression) was consistently lower(range: 3.2% [Georgia]10.9% [Missouri]) (able 4).

Cmparisn Beween 2002 and 2006Prevalence Esimaes

All o the 10 ADDM sites that provided data or both2002 and 2006 reported an increase in identiied ASDprevalence among children aged 8 years within each site(range: 34%95%; p


13/28


TABLE 4. Median age in months o the earliest dignosis o an autism spectrum disorder (ASD) with developmental concerns* aage 3 years, by median age in months at which concerns were noted, and proportion and median age in months o children owhom developmental regression or plateau was noted in records Autism and Developmental Disabilities Monitoring Network11 Sites, United States, 2006

Site

Totalno. withASDs

Age o earliestASD diagnosis

on recordGeneralconcern Social Language

Imaginativeplay

Developmentalregression

Developmentaplateau

Median (Range) % Median % Median % Median % Median % Median % Median

Alabama 212 51 (2299) 95.3 24 mos 55.7 36 mos 85.8 24 mos 24.5 36 mos 23.1 24 4.7 21Arizona 504 59 (5104) 88.3 24 mos 42.5 36 mos 81.5 24 mos 17.1 36 mos 13.3 24 5.8 24

Colorado 54 60 (2397) 85.2 24 mos 48.1 36 mos 75.9 24 mos 25.9 36 mos 29.6 30 3.7 23Florida 116 41 (998) 94.0 12 mos 27.6 24 mos 84.5 12 mos 6.9 24 mos 28.4 18 3.4 15Georgia 474 53 (2101) 90.7 24 mos 41.4 36 mos 79.7 24 mos 13.1 36 mos 20.7 19 3.2 24

Maryland 243 58 (18105) 69.5 24 mos 36.6 36 mos 54.7 24 mos 7.8 36 mos 23.0 18 3.7 24

Missouri 321 53 (10102) 81.3 24 mos 43.3 24 mos 72.6 24 mos 11.8 36 mos 24.3 18 10.9 24North

Carolina230 50 (15106) 86.5 24 mos 47.4 36 mos 75.2 24 mos 23.5 36 mos 23.9 19 8.7 21

Pennsylvania 150 52 (19105) 92.0 24 mos 46.0 36 mos 82.0 24 mos 16.7 36 mos 22.0 24 3.3 12South

Carolina196 54 (13103) 89.3 24 mos 37.8 36 mos 83.2 24 mos 16.3 36 mos 21.4 15 6.1 18

Wisconsin 257 53 (18106) 87.9 12 mos 50.2 24 mos 77.0 24 mos 31.5 24 mos 29.6 18 7.4 18

* For each child, all evaluation inormation was reviewed to categorize any concerns noted in developmental evaluations concerning the childs developmental status beore age 3 years; any specifed concerns regarding the development o social, language, or imaginative play beore age 3 years also weredocumented.

Indicates that the developmental concern was noted to occur at or beore age either 12, 24, or 36 months.

still resided in the surveillance areas did not change >5%. Tethree sites (Arizona, Missouri, and North Carolina) that hadthe greatest increases in ASD prevalence also had the largestincreases during 20022006 in the proportion o children aged8 years who were born in and still resided in the surveillancearea, indicating a more stable population by 2006. Some dier-

FIGURE 3. Intelligence quotient (IQ) o children aged 8 years with an autism spectrum disorder (ASD) or whom psychometrictest data were available,* by site, sex, and IQ score Autism and Developmental Disabilities Monitoring Network, 11 sites, United

States, 2006

ences were noted in the proportion o les that were located orreview. Sensitivity analyses conducted by all sites determinedthat les that were eligible or review but not located contributed to 85)

BorderlineIQ (7185)

CognitivelyImpaired(70)

Male Female

Alabama

Male Male Male Male MaleFemale Female Female Female Female

Arizona Colorado

Site

Percen

tage

Georgia North Carolina South Carolina

*Sites with psychometric test data on 70% o children identifed with an ASD were included.


14/28


15/28


FIGURE 4. Change in prevalence* o autism spectrum disorders (ASDs) among children aged 8 years, by classifcation type Autism and Developmental Disabilities Monitoring Network (ADDM), 10 sites, United States, 2002 and 2006

Site

Alabama Arizona Colorado Georgia Mar yland Missouri Nor thCarolina

Pennsylvania SouthCarolina

Wisconsin

Percentage

0

2

4

6

8

10

12

14

ADDM ASD prevalence

02 02 02 02 02 02 02 02 02 02

Previously classified ASD prevalence

06 06 06 06 06 06 06 06 06 06

* Per 1,000 children aged 8 years in the surveillance area. ADDM ASD prevalence is indicated by the entire bar.

FIGURE 5. Change in subtype o autism spectrum disorder (ASD) as documented in evaluation records, by year Autism andDevelopmental Disabilities Monitoring Network, 10 sites, United States, 2002 and 2006

Site

Alabama Arizona Colorado Georgia Maryland Missouri NorthCarolina

Pennsylvania SouthCarolina

Wisconsin

Percen

tage

0

10

20

30

40

50

60

70

80

90

100

02 06 02 06 02 06 02 06 02 06 02 06 02 06 02 06 02 06 02 06

% with nodocumentedASD diagnosis

% diagnosedwith ASD* butnot autism

% diagnosedwith autisticdisorder ever

* Includes documented diagnoses o Asperger disorder, Pervasive Developmental DisorderNot Otherwise Specifed, Autism Spectrum Disorder or PervasiveDevelopmental Disorder, or an International Classifcation o Diseases, 9th Revision, Clinical Modifcation (ICD-9-CM) code o 299.8.

prevalence increase rom 2002 to 2006 could be explained byimproved ability to locate records or review.

For the 10 sites that reported data or both 2002 and 2006,an average o our evaluations were abstracted per child in2002 compared with ve evaluations abstracted per child in

2006. In addition, a qualitative assessment o dierences inthe inormation contained in developmental evaluation recordsacross sites conducted by polling the clinicians reviewing therecords indicated that the quality and amount o inormationcontained in the evaluation records improved steadily over


16/28


TABLE 6. Number o children aged 8 years with autism spectrum disorder (ASD), by median age in months at earliest knowndiagnosis Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2002 and 2006

Site

2002 2006

Change in age oearliest ASD diag-nosis on record,

2002 to 2006

Age* o earliest ASD diagnosis on record Age o earliest ASD diagnosis on record

Total no. withASDs Median Range

Total no. withASDs Median Range

Alabama 116 66 (10101) 212 51 (2299) -15Arizona 280 63 (20101) 504 59 (5104) -4

Colorado 65 62 (12100) 54 60 (2397) -2Georgia 337 58 (23103) 474 53 (2101) -5Maryland 199 60 (21105) 243 58 (18105) -2

Missouri 205 56 (20106) 321 53 (10102) -3North Carolina 135 53 (2199) 230 50 (15106) -3

Pennsylvania 111 58 (2494) 150 52 (19105) -6South Carolina 140 60 (14103) 196 54 (13103) -6

Wisconsin 181 54 (11104) 257 53 (18106) -1

* In months Data published previously or 2002 (CDC. Prevalence o autism spectrum disordersautism and developmental disabilities monitoring network, 14 sites,

United States, 2002. In: Surveillance Summaries, February 9, 2007. MMWR 2007;56[No. SS-1]:1228) reported South Carolinas earliest age o ASDdiagnosis at age 64 months. These data have been updated to include the earliest known ASD diagnosis reported in the records.

time, with noticeable increases in level o detail documentedon ASD symptoms by 2006.

DiscussinAcross the 11 ADDM sites, retrospective estimates rom

the year 2006 indicated ASD prevalence ranging rom 4.2 to12.1, with most sites identiying prevalence o 7.5 to 10.4 casesper 1,000 children aged 8 years (average: 9 cases). Tis reportprovides updated data rom the ADDM Network indicatingan average prevalence o ASDs in the United States approach-ing 1%, which is substantially higher than previously reportedADDM average estimates (4,12). Te rst ADDM Networkreports rom 2000 and 2002 indicated average identied ASDprevalence among children aged 8 years o 6.7 and 6.6 cases per1,000 children, respectively, with one site, New Jersey, identiy-ing prevalence o approximately 10 cases per 1,000 children in2000 and 2002 (4,12). A recent U.S. study analyzing nationalsurvey data or 2007 o parents reports o ASDs indicated thatapproximately 1% o all children aged 317 years and 1.3%o all children aged 68 years had an ASD (21). Recentlypublished international studies have reported ASD prevalenceestimates upward o 1% (12,1720). Recent studies indicating

changes in ASD prevalence are consistent in nding increases inidentied ASD prevalence. However, no studies have addressedthe multiple complex contributions o changes in ascertain-ment and potential risk actors. Because ASDs are behaviorallydened conditions identied on the basis o observation andreported symptoms, evaluations o ASD prevalence acrossmultiple studies and methodologies are especially inormative.Te ADDM Network provides a systematic conrmation odocumented ASD classications and symptoms. In addition,

the methodology identies children who have documentedsymptoms o an ASD but have not previously received an ASDdiagnosis by a community proessional, increasing complete-ness o ASD prevalence estimates.

In general, ADDM ASD prevalence estimates were lower insites that relied solely on health sources to identiy cases com-pared with sites that also had access to education records. InMissouri, although investigators did not have direct access tospecial education evaluation records, this inormation could beound in the regional developmental disabilities programs, whichprovide evaluation and treatment coordination. Lack o accessto education records appears to have had the greatest impact on

identied ASD prevalence in Alabama and Florida.Consistent with ndings rom previous ADDM Network

reports and other studies (24,12), data rom the 2006 surveillance year indicated that ASD prevalence was 4.5 times higherin males than in emales. Among children aged 8 years, ASDprevalence was 14.5 per 1,000 males compared with 3.2 per1,000 emales. Research exploring these marked sex dier-ences suggests dierences in the developmental proles andpotential etiologies between males and emales with ASDs (33)Identied ASD prevalence also varied by race/ethnicity, withsome sites showing higher ASD prevalence estimates amongnon-Hispanic white children than among non-Hispanic blackchildren and Hispanic children. Considerable variability hasbeen reported across studies in the racial disparities o ASDs(33). Further examination o socioeconomic actors and sub-groups o children might provide more clarity on whether raciaor ethnic disparities refect dierences in ascertainment issuesenvironmental risk actors, and genetic susceptibility.

Children identied with an ASD in 2006 refect a group withless co-occurring cognitive impairment than the population


17/28


18/28


sis (range: 1 month [Wisconsin]15 months [Alabama]), withan overall trend across all sites o a reduction o approximately5 months. Although encouraging, this trend toward earlieridentication was modest and still quite delayed comparedwith the documented concerns about the childs developmentoccurring beore age 2 years. For children identied with an

ASD by age 8 years, the reduction o age o identication doesnot appear to be a substantial actor infuencing prevalenceestimates; however, urther analysis o this issue is needed.

he variation in ASD prevalence among demographicsubgroups might refect local patterns in who is evaluated ordevelopmental concerns, in how those concerns and behaviorsare documented in developmental evaluation records, and indierential risk or ASD according to these subgroups. Teincrease in ASD prevalence among males is not unexpectedgiven the consistently reported sex dierence, but whetherthis is attributable to improved identication or to increasedrisk among some males is not known. Te greater variabilityamong increases or emales is more likely to be attributableto improved recognition o ASD symptoms among emales.Dierences in identied ASD prevalence by race/ethnicityhave been hypothesized to refect changes in identicationpatterns because no clear etiologic hypotheses have been pro-posed that would predict dierences in ASD prevalence byrace or ethnicity. Te role o race/ethnicity is not independento socioeconomic actors related to community identicationpatterns (33,46,47). Te role o these actors as risk indicatorsis debated (33).

Factors related to the record review process that might aect

identied ASD prevalence include the type and stability o datasources, screening criteria, and ability to locate les or reviewover time. Data sources were relatively stable or all sites, andalthough sites with access to both health and education datahad higher average identied ASD prevalence, the inclusion oeducation sources was consistent or these sites rom 2002 to2006. Application o screening criteria also was stable withinsites rom 2002 to 2006. For certain sites, the ability to locaterecords between the two time periods diered, but this ac-tor explained


19/28


Srenghs he ADDM NewrkMehdlgy

Te systematic implementation o the ADDM Networkmethodology, which requires standardized training o abstrac-tors and clinician reviewers, ongoing monitoring or quality

assurance, and standardization o methods to conrm casesand conduct analyses, produces highly reliable and valid data(27). Te level o detail collected and reviewed on childrenrom multiple sources improves on previous estimates basedsolely on administrative records or single-source surveillance.Te use o consistent methods across sites and over time hasenabled researchers to begin to evaluate changes in ASDprevalence. Te ADDM methods have great utility given theirability to identiy children with the prole consistent with anASD regardless o whether the child was classied as havingautism or services. On the basis o ADDM data, i prevalenceestimates depended solely on the documented classication o

ASD, prevalence would be underestimated by 5%22%. Teability to link ADDM data with external data sources aordsthe ability to examine potential risk actors beyond the scopeo surveillance data alone. Routine linkages o ADDM datawith birth certicate and census les enables evaluation osuch potential risk actors as parental age (48), multiple births(49), and socioeconomic characteristics (33,47). In addition,determining the prevalence o children at age 8 years reducesthe infuence o earlier ages o identication and enables theinclusion o children whose developmental issues might notbe recognized until they reach school-age (37).

LimiainsTe data provided in this report are subject to at least three

limitations. First, multiple actors can contribute to changesin prevalence estimates obtained through retrospective recordreview (27). Te existence o the records, the ability to locatethem, and the quantity and quality o the inormation in therecords are all relevant issues that aect ASD prevalence usingADDM methods. Because overall prevalence was lower amongthe sites with access to health evaluations alone, sites that lackedaccess to education evaluations or to another source that pro-vides ree evaluations to the public likely underestimated ASD

prevalence. In addition, sites without access to education evalu-ations have been unable to obtain sucient data to examinethe role o cognitive unctioning among children with ASDs.Tese points underscore the need to improve the accessibilityo education sources and evaluations. Second, the majorityo sites did not include private schools, charter schools, andclinical providers or service centers with small numbers o cli-ents. Previous reports suggest that the impact o not reviewingrecords o children who are private or homeschooled is mini-

mal (27). Regardless o the type o data source access, specicdata sources rom 2002 to 2006 were relatively stable withinall sites. Finally, direct in-person evaluation o each child todetermine case status is not part o a records-based approach tosurveillance. However, the surveillance approach minimizes theburden on children and their amilies and reduces response bias

and validation o the symptom prole consistent with an ASDis veried by clinician review. In addition, CDC is conductinga validation study comparing records-based to direct evaluationmethods. Although the ADDM sites were not selected to bea nationally representative sample, the population includedrepresents a substantial number o children, >300,000 (7.9%)o all children in the United States aged 8 years in 2006.

Public Healh ImplicainsTe data in this report, which indicate prevalence o ASDs

approaching approximately 1% o U.S. children aged 8 years,

corroborate other recently published data (12,1721)and conrmthat ASDs are an important public health concern. Te progressiveincreases in ASD prevalence recorded during 20022006 urtherunderscore the need to understand better the risk actors andcauses o these conditions. Although researchers can begin to evaluate trends with the data provided in this report, caution is urgedgiven the likelihood o some variation over time in prevalence obehaviorally dened conditions such as ASDs. No single actorexplains the changes identied in ASD prevalence over time, andmuch needs to be done to understand the relative contribution othe multiple actors involved. Although some o these increasescan be accounted or by improved identication and awareness

the steady increase in ASD symptoms in the population possiblyrefects increased risk, particularly among males. Some progresshas been reported in quantiying the eect o single actors suchas reduction in age o diagnosis and inclusion o milder cases(45) or shiting in diagnostic patterns (50) on increased autismprevalence. However, these analyses are limited by the use osingle-source datasets, which contain primarily data on childrenwith autism rather than on those with the whole spectrum oASDs. More complete datasets and complex models are neededto assess urther the infuence o the multiple actors discussed inthis report on the changing prevalence o ASDs.

Whether identied ASD prevalence estimates will plateau orcontinue to increase is unknown. Te ADDM cohorts in thisreport comprise children born in 1994 (or 2002 data) and1998 (or 2006 data). Children born starting in the mid-to-late1990s were particularly susceptible to the changing infuenceo the new DSM-IV criteria in 1994 and to increased autismawareness among the public and health proessionals (2,3)Te impact on estimated ASD prevalence o the broadeningo diagnostic criteria and the increased awareness o ASDs


20/28


might reach a high point and then diminish ater a period otime. Tereore, evaluating the prevalence o ASDs amongchildren born in this millennium using the same standard orevaluating the change in ASD symptoms over time is criticalto understanding the current and changing population ochildren with ASDs. O note, recent research indicates that

the core social traits o autism are distributed in the populationalong a continuum(51); where the line is drawn between traitmeasures regarded as normal variance in behavior versus thoselabeled as impairment or disability might aect ASD prevalenceestimates. Te landscape o ASD diagnoses is likely to changewith the introduction o the next version o the DSM expectedin 2012. Eorts are needed to examine prevalence changes inother childhood conditions such as attention-decit/hyperac-tivity disorder, asthma, and allergies to assess whether changesin ASD prevalence are occurring in isolation (5254).

Eorts are needed to understand how complex genetic andenvironmental actors interact to result in the symptoms whichmake up the autism spectrum. In addition to changes in ASDprevalence by race/ethnicity, sex, and cognitive unctioning,other potential risk actors (e.g., variations by urban and ruralarea, sociodemographic status, perinatal complications, andparental age) also need to be studied urther. ADDM dataare being analyzed to understand better the roles o theseand other actors. Studies such as the Study to Explore EarlyDevelopment, a CDC-unded study examining a wide arrayo risk actors or ASD are being conducted and are necessaryto test hypotheses more ully. In addition, the coordinationo research priorities between public and private organizations

through the Interagency Autism Coordinating Committee othe National Institutes o Health and acceleration o researchon ASDs highlights the need or an urgent, coordinated, andmultiprong approach to ASD research.

Experienced clinicians using standardized methods candiagnose autism reliably in children as young as age 2 years,and these diagnoses tend to be stable within the ASD spectrum(55,56). Tese data provide urther evidence (57) that a signi-cant lag exists between earliest concerns and actual reportedidentication o an ASD, thus contributing to potentiallysignicant delays in intervention. Because o the benet oearly intervention (40), identication o ASDs at earlier ages

is essential to ensure that children in the United States receiveoptimal early intervention services. In addition to communityproessionals providing diagnosis, many children who did nothave an ASD diagnosis recorded in their record as o 2006were receiving special education services through an autismeligibility in the public schools. Public schools are playing acrucial role in evaluating, identiying, and serving children

with ASDs. CDC is working with caregiver and proessionalgroups through the Learn the Signs. Act Early. public aware-ness campaign to provide education on the early recognitiono signs o ASDs and other developmental disabilities (41). Inaddition, in 2007, the American Academy o Pediatrics recom-mended that routine screening or autism occur or all children

as part o ongoing developmental screening during the 18- and24-month well-child visits (42,43). In the uture, ADDM datacan be used to monitor changes in early identication thatpotentially result rom increased education eorts.

More children than ever beore are receiving services orASDs and are having symptoms o ASDs documented indevelopmental evaluation records. Even without ully under-standing the complex causes o this increase in identied ASDprevalence, the impact on aected children, amilies, andcommunities is substantial. Prevalence estimates can be usedto plan policy, educational, and intervention services needs orpersons with ASDs. In addition to continued evaluation oASD prevalence changes, major collaborative eorts are neededto improve research into what actors put certain people at riskand how to intervene to help reduce the debilitating symptomso ASDs. Concerted eorts are essential to address the manyneeds o aected persons and to provide coordinated supportservices which improve daily unctioning and long-term lieoutcomes.

AcknwledgmensTe Autism and Developmental Disabilities Monitoring (ADDM)

Network projects were unded by CDC. Additional inormationabout these projects is available at http://www.cdc.gov/autism .

Inormation in this report was provided by ADDM NetworkSurveillance Year 2004 and 2006 principal investigators: BeverlyMulvihill, PhD, Martha Wingate, PhD, University o AlabamaBirmingham; Russell S. Kirby, PhD, University o South FloridaSydney Pettygrove, PhD, Chris Cunni, MD, F. John MeaneyPhD, University o Arizona, ucson; Lisa Miller, MD, ColoradoDepartment o Public Health and Environment, Denver; CordeliaRobinson, PhD, University o Colorado at Denver and HealthSciences Center, Denver; Gina Quintana, Colorado Departmeno Education; Marygrace Yale Kaiser, PhD, University o MiamiCoral Gables, Florida; Li-Ching Lee, PhD, Johns HopkinUniversity, Baltimore, Maryland; Rebecca Landa, PhD, Kennedy

Krieger Institute, Baltimore, Maryland; Craig Newschaer, PhDDrexel University, Philadelphia, Pennsylvania; John ConstantinoMD, Robert Fitzgerald, MPH, Washington University in St. LouisMissouri; Julie Daniels, PhD, University o North Carolina, ChapeHill; Ellen Giarelli, EdD, Jennier Pinto-Martin, PhD, Universityo Pennsylvania, Philadelphia: Susan E. Levy, MD, Te ChildrenHospital o Philadelphia, Pennsylvania; Jane Charles, MD, JoyceNicholas, PhD, Medical University o South Carolina, CharlestonMaureen Durkin, PhD, DrPH, University o Wisconsin, MadisonCatherine Rice, PhD, Jon Baio, EdS, Kim Van Naarden Braun, PhD
http://www.cdc.gov/autismhttp://www.cdc.gov/autism


21/28


22/28


29. World Health Organization. International statistical classication o diseasesand related health problems, ninth revision, clinical modication, ICD-9-CM.4th ed. Geneva, Switzerland: World Health Organization; 1997.

30. Selvin S. Statistical power and sample-size calculations. In: Selvin S, ed.Statistical analyses o epidemiologic data. 2nd ed. New York, NY: OxordUniversity Press; 1996.

31. SAS, Inc. SAS or Windows, Rel. 9.1. Cary, NC: SAS Institute Inc.; 2004.32. SPSS, Inc. SPSS or Windows, Rel. 13. Chicago, IL: SPSS Inc.; 2005.33. Newschaer C, Croen LA, Daniels J, et al. Te epidemiology o the autism

spectrum disorders. Annu Rev o Public Health 2007;28:23558.34. Chakrabarti S, Fombonne E. Pervasive developmental disorders in

preschool children. JAMA 2001;285:30939.35. Losh M, Adolphs R, Poe MD, et al. Neuropsychological prole o autism

and the broad autism phenotype. Arch Gen Psychiatry 2009;66:51826.36. Lord C, Shulman C, DiLavore P. Regression and word loss in autistic

spectrum disorders. J Child Psychol and Psychiatry 2004;45:93655.37. Fombonne E. Te prevalence o autism. JAMA 2003;289:879.38. Charman . Te prevalence o the autism spectrum disorders: recent evidence

and uture challenges. Eur Child Adoles Psychiatry 2002;11:24956.39. Posserud M, Lundervold AJ, Lie SA, Gillberg C. Te prevalence o autism

spectrum disorders: impact o diagnostic instrument and non-response bias.Soc Psychiatry Psychiatr Epidemiol 2009. [Epub]. Available at http://www.springerlink.com/content/650012524603240p.Accessed December 7, 2009.

40. National Research Council, Committee on Educational Interventions orChildren with Autism, Division o Behavioral and Social Sciences andEducation. Educating children with autism. Washington, DC: National

Academy Press; 2001.41. CDC. Learn the signs. Act early. Atlanta, GA: US Department o Health

and Human Services, CDC; 2006. Available at http://www.cdc.gov/actearly. Accessed December 7, 2009.

42. American Academy o Pediatrics Council on Children with Disabilities.Identiying inants and young children with developmental disordersin the medical home: an algorithm or developmental surveillance andscreening. Pediatrics 2006;118:40520.

43. Johnson CP, Myers SM; American Academy o Pediatrics Council onChildren With Disabilities. Identication and evaluation o children

with autism spectrum disorders. Pediatrics 2007;120:1183215.44. Parner E, Schendel D, Torsen P. Autism prevalence trends over time

in Denmark: changes in prevalence and age at diagnosis. Arch PediatrAdolesc Med 2008;162:11506.45. Hertz-Picciotto I, Delwiche L. Te rise in autism and the role o age at

diagnosis. Epidemiology 2009;20:8490.

46. Mandell DS, Listerud J, Levy S, Pinto-Martin J. Race dierences inthe age o diagnosis among Medicaid-eligible children with autism

J Am Acad Child Adolesc Psychiatry 2002;41:144753.47. Mandell DS, Wiggins LD, Carpenter LA, et al. Racial/ethnic dispari

ties in the identication o children with autism spectrum disordersAm J Public Health. 2009;99:4938.

48. Durkin MS, Maenner MJ, Newschaer CJ, et al.Advanced parentalage and the risk o autism spectrum disorder. Am J Epidemio2008;168:126876.

49. Van Naarden Braun K, Schieve L, Daniels J, et al. Relationshipbetween multiple births and autism spectrum disorders, cerebral palsyand intellectual disabilities: Autism and Developmental DisabilitieMonitoring (ADDM) Network2002 Surveillance Year. Autism Re2008;1:265316.

50. King M, Bearman P. Diagnostic change and the increased prevalence oautism. Int J Epidemiol 200938:122434.

51. Constantino JN, odd RD. Autistic traits in the general population: twin study.Arch Gen Psychiatry2003;60:52430.

52. Atladttir H, Parner E, Schendel D, Dalsgaard S, homsen PTorsen P. ime trends in reported diagnoses o childhood neuropsychiatric disorders: a Danish cohort study. Arch Pediatr Adolesc Med2007;161:1938.

53. Woodru J, Axelrad DA, Kyle AD, Nweke O, Miller GG, Hurley

BJ. rends in environmentally related childhood illnesses. Pediatrics2004;113:113340.

54. Pallapies, D. rends in childhood disease. Mutat Res 2006;608:1001155. Lord C, Risi S, DiLavore PS, Shulman C, Turm A, Pickles A. Autism

rom 2 to 9 years o age. Arch Gen Psychiatry 2006;63:694701.56.Van Daalen E, Kemner C, Dietz C, Swinkels SH, Buitelaa

JK, Van Engeland H. Inter-rater reliability and stability o diagnoses oautism spectrum disorder in children identied through screening at very young age. Eur Child Adolesc Psychiatry2009. [Epub]. Availableat http://www.springerlink.com/content/b57406752324wjk6. AccessedDecember 7, 2009.

57. Wiggins L, Baio J, Rice C. Examination o the time between rst evaluation and rst autism spectrum diagnosis in a population-based sample

J Dev Behav Pediatr 2006;27(2 Suppl):S7987.
http://www.springerlink.com/content/650012524603240phttp://www.springerlink.com/content/650012524603240phttp://www.cdc.gov/actearlyhttp://www.cdc.gov/actearlyhttp://www.springerlink.com/content/b57406752324wjk6http://www.springerlink.com/content/b57406752324wjk6http://www.cdc.gov/actearlyhttp://www.cdc.gov/actearlyhttp://www.springerlink.com/content/650012524603240phttp://www.springerlink.com/content/650012524603240p


23/28


Appendix

Brie Updae: Prevalence Auism Specrum Disrders (ASDs) Auism and Develpmenal Disabiliies Mniring (ADDM) Newrk,

Unied Saes, 2004

CDC and seven site project teams rom the Autism andDevelopmental Disabilities Monitoring (ADDM) Network(1,2) (able 1) collaborated in monitoring the prevalence oautism spectrum disorders (ASDs) in 2004 in selected areaso eight states: Alabama (three counties in central Alabama), Arizona (districts in one county, including metropolitanPhoenix), Georgia (the CDC site in ve counties in metropoli-tan Atlanta), Maryland (ve counties in suburban Baltimore),Missouri (ve counties in metropolitan St. Louis), NorthCarolina (eight central counties), South Carolina (23 coun-ties in the Coastal and PeeDee regions), and Wisconsin (three

counties in south-central Wisconsin).Te 2004 surveillance year represented a smaller-scale eortthan other ADDM Network surveillance years conductedpreviously (3,4) and in 2006. Participation in the 2004 surveil-lance year was undertaken by sites that had the available timeand resources to collect additional data and it included a smallercatchment area than in 2006 or ve o the eight sites (Alabama,Arizona, Maryland, North Carolina, and Wisconsin). Becausethe smaller population monitored in 2004 might not be com-parable to the larger area surveyed in other years, caution isneeded when comparing 2004 results with those rom othersurveillance years. However, these additional data provide

inormation on the status o identied ASD prevalence in thesites included in this summary.

MehdsPrevalence o ASDs in U.S. children was estimated through

a systematic retrospective review o evaluation records inmultiple sites participating in the ADDM Network. Dataor 2004 were collected retrospectively rom existing recordsrom eight ADDM Network sites. Children aged 8 years (i.e.,those born in 1996) with a notation o an ASD or descriptionsconsistent with an ASD were identied through screening and

abstraction o existing health and education records containingreported proessional assessments o the childs developmentalprogress at health-care or education acilities. Children aged 8years whose parent(s) or legal guardian(s) resided in the respec-tive areas in 2006 met the case denition or an ASD i theirrecords documented behaviors consistent with the Diagnosticand Statistical Manual o Mental Disorders, 4th edition, textrevision (DSM-IV-R) criteria or autistic disorder, pervasive

developmental disorder, not otherwise specied (PDD NOS)or Asperger disorder. Presence o an identied ASD was deter-mined through a review o data abstracted rom developmentaevaluation records by trained clinician reviewers.

For 2004, a limited (or streamlined) abstraction and reviewo evaluation records was conducted by six o the eight sites(all except North Carolina and South Carolina) or childrenwith a previously documented classication o an ASD (i.e., adiagnosis o an ASD or a special education eligibility classica-tion o autism documented in the records). Tis limited reviewwas based on data obtained rom MADDSP ASD surveillance

and indicated that 98.0% o children aged 310 years with aprevious ASD diagnosis and 99.0% o children with a previ-ous autism eligibility or special education services satised thesurveillance criteria or having an ASD (5). Tis addition wasimplemented to improve eciency without altering the basiccase classication standards. Te limited abstraction includedall inormation as noted above except the verbatim behaviordescriptions. CDCs analysis o the resource savings indicatedthat they were not worth the loss o the variables summarizingthe behavioral descriptions collected. Tereore, or the 2006surveillance year, the standard abstraction and review method-ology was used instead o the limited review used in 2004.

For 2004 surveillance year records that underwent limitedabstraction, children were considered to meet the ASD casedenition on the basis o the previously documented ASDdiagnosis unless 1) conficting inormation was noted in therecord, 2) the reviewer needed additional inormation, or 3) therecord indicated that an ASD had been ruled out as a diagnosisIn those circumstances, a ull abstraction was perormed, andthe case was reviewed again by the clinician reviewer. For anyrecord, i a child met the ASD case denition, and the clinicianreviewer had cause to question the case status, the reason wasnoted, and a blinded secondary review was undertaken. FinaASD case status was established on the basis o a consensus

review. For all sites, the range o agreement or nal case de-nition was acceptable (76.0%94.0%; Kappa range 0.50.9or 2004. Methods or 2004 were otherwise comparable tothose used or the 2006 surveillance year. Prevalence resultsare reported per 1,000 children aged 8 years. Chi-square testswere used to compare prevalence estimates within and acrosssites.


24/28


TABLE 1. Number* and percentage o children aged 8 years, by race/ethnicity and site Autism and Developmental DisabilitiesMonitoring Network, eight Sites, United States, 2004

SiteSite

institution Surveillance areaTotalno.

White, non-Hispanic

Blac, non-Hispanic Hispanic API AI/AN

Receivingspecial

education(%))

Abstractedor ASD**

(%)No. (%) No. (% ) No. (% ) No. (% ) No. (% )

Sites with access to health records

Alabama Univ o AlabamaBirmingham 3 counties in centralAlabama 11,676 6,765 (57.9) 4,329 (37.1) 377 (3.2) 166 (1.4) 39 (0.3) (16.7) (1.6)

Missouri WashingtonUnivSt. Louis

5 counties includingmetropolitan St. Louis

26,970 18,818 (69.8) 6,651 (24.7) 728 (2.7) 697 (2.6) 76 (0.3) (19.0) (1.1)

Wisconsin Univ o WisconsinMadison

3 counties in southcentral Wisconsin

11,312 9,593 (84.8) 554 (4.9) 634 (5.6) 497 (4.4) 34 (0.3) (14.4) (1.3)

Sites with access to education and health records

Arizona Univ oArizonaTucson

1 county (Maricopa) inmetropolitan Phoenix

13,620 6,571 (48.2) 713 (5.2) 5,576 (40.9) 422 (3.1) 337 (2.5) (10.2) (2.1)

Georgia CDC 5 counties includingmetropolitan Atlanta

45,190 18,270 (40.4) 19,176 (42.4) 5,167 (11.4) 2,461 (5.4) 116 (0.3) (10.8) (1.7)

Maryland Johns HopkinsUniv

5 counties in suburbanBaltimore

20,981 15,044 (71.7) 4,213 (20.1) 685 (3.3) 976 (4.7) 63 (0.3) (12.8) (1.3)

North Carolina Univ o NorthCarolinaChapelHill

8 counties in centralNorth Carolina

20,187 11,670 (57.8) 5,798 (28.7) 2,070 (10.3) 570 (2.8) 79 (0.4) (14.4) (1.6)

South Carolina Medical Univ oSouth Carolina

23 counties in theCoastal and Pee Dee

regions

22,399 11,875 (53.0) 9,335 (41.7) 824 (3.7) 256 (1.1) 109 (0.5) (16.2) (1.1)

* Total numbers o children aged 8 years in each study area obtained rom CDCs National Center or Health Statistics (NCHS) vintage 2007 postcensal population estimatesSurveillance area denominators exclude those school districts that did not allow access to records.

Asian/Pacifc Islander.American Indian/Alaska Native.Number o students in special education and total enrolled in districts in surveillance area or the 20032004 school year obtained rom http://www.nces.ed.gov.**Autism spectrum disorder. Represents the number o children identifed as possibly having an ASD divided by the total number o children aged 8 years in the population.

ResulsOn the basis o postcensal estimates, the number o children

aged 8 years in the eight surveillance sites ranged rom 11,312in Wisconsin to 45,190 in Georgia (able 1) (6). Distributionaccording to race or ethnicity among children aged 8 years

varied across surveillance sites. Te percentage o non-Hispanicwhite children residing in each surveillance area ranged rom40.4% in Georgia to 84.8% in Wisconsin; the percentage onon-Hispanic black children ranged rom 4.9% in Wisconsinto 42.4% in Georgia; the percentage o Hispanic childrenranged rom 2.7% in Missouri to 40.9% in Arizona; the per-centage o Asian/Pacic Islander children ranged rom 1.1%in South Carolina to 5.4% in Georgia; and the percentage oAmerican Indian/Alaska native children ranged rom 0.5%in all sites except Arizona, with 2.5%. Although most siteshad a similar distribution o the population by race/ethnicity,

compared with the larger surveillance areas or 2006, Alabamassmaller 2004 surveillance area included a greater proportiono non-Hispanic black children, and Wisconsins 2004 areaincluded more non-Hispanic white children (able 1). Tebreakdown by sex was similar across sites, with approximatelyequal distribution o male and emale children (able 1).

In 2004, the overall identied ASD prevalence per 1,000children aged 8 years varied across ADDM sites (range: 4.6[Alabama] 9.8 [Arizona]) (able 2). Te average across

all sites was 8.0 (CI = 7.68.4) per 1,000 children. Amongthe eight 2004 sites, six were clustered in a tighter range(7.89.8 per 1,000 children), and these rates did not dierrom each other signicantly. However, prevalence in Alabama(4.6) and South Carolina (5.3) was signicantly (p


25/28


TABLE 2. Estimated prevalence* (prev) o autism spectrum disorders (ASDs) among children aged 8 years, by sex and race/ethnicity Autism and Developmental Disabilities Monitoring Network, 8 Sites, United States, 2004

SiteTotal no.chldren

Totalno. withASDs

Sex

Male-to-emaleprev ratio

Total Males Females

Prev 95% CI Prev 95% CI Prev 95% CI


Alabama 11,676 54 4.6 (3.56.0) 6.8 (5.09.3) 2.3 (1.33.9) 3.0Missouri 26,970 221 8.2 (7.29.4) 13.5 (11.715.6) 2.8 (2.03.8) 4.9Wisconsin 11,312 88 7.8 (6.39.6) 12.0 (9.415.2) 3.7 (2.45.6) 3.3

Sites with access to education and health recordsArizona 13,620 133 9.8 (8.211.6) 15.8 (13.119.1) 3.6 (2.45.3) 4.4

Georgia 45,190 401 8.9 (8.19.8) 14.1 (12.615.7) 3.6 (2.84.4) 4.0Maryland 20,981 185 8.8 (7.610.2) 14.1 (12.016.5) 3.2 (2.34.6) 4.4

North Carolina 20,187 176 8.7 (7.510.1) 14.8 (12.617.3) 2.4 (1.63.6) 6.1South Carolina 22,399 118 5.3 (4.46.3) 8.9 (7.410.8) 1.5 (0.92.4) 6.1

Race/ethnicity Prev ratio

Site

White, non-Hispanic Black, non-Hispanic Hispanic API**White-to-

blackWhite-to-Hispanic

Black-to-HispanicPrev 95% CI Prev 95% CI Prev 95% CI Prev 95% CI


Alabama 3.8 (2.65.6) 6.0 (4.18.8)

0.6 Missouri 8.7 (7.510.2) 3.2 (2.14.8) 5.5 (2.114.6) 4.3 (1.413.4) 2.8 1.6 0.6Wisconsin 7.4 (5.99.3) 3.6 (0.914.4) 1.6 (0.211.2) 6.0 (2.018.7) 2.0 4.7 2.3

Sites with access to education and health recordsArizona 12.6 (10.215.7) 5.6 (2.115.0) 7.0 (5.19.6) 11.9 (4.928.5) 2.3 1.8 0.8

Georgia 9.7 (8.411.3) 7.9 (6.89.3) 6.4 (4.59.0) 8.1 (5.212.6) 1.2 1.5 1.2Maryland 7.4 (6.29.0) 12.8 (9.816.7) 8.8 (3.919.5) 12.3 (7.021.7) 0.6 0.8 1.5

North Carolina 8.6 (7.010.4) 9.0 (6.811.8) 6.8 (4.011.4) 5.3 (1.716.3) 1.0 1.3 1.3South Carolina 5.5 (4.37.0) 4.1 (3.05.6) 2.4 (0.69.7) 1.3 2.3 1.7

* Per 1,000 children aged 8 years. All children are included in the total regardless o race or ethnicity. The total also includes children or whom race/ethnicity was unknown. All male-to-emale ratios dierent within sites (p


26/28


children identied also had cognitive impairment. Variabilityexisted by race and ethnicity, with some sites showing higherASD prevalence estimates among non-Hispanic white childrenthan non-Hispanic black children and Hispanic children.

Considering data rom the 2004 surveillance year illustratesthe incremental increases in identied ASD prevalence over the

short time period 20022006 based on the ADDM Networkmethods. For all sites except South Carolina, where ASDprevalence decreased rom 2002 to 2004 and then increasedconsiderably rom 2004 to 2006, the data indicate the incre-mental and steady increase in identied ASD prevalence overthe period reported.

Reerences1. Rice CE, Baio J, Van Naarden Braun K, Doernberg N, Meaney F J, Kirby

RS, or the ADDM Network. A public health collaboration or thesurveillance o autism spectrum disorders. Paediatr Perinat Epidemiol2007;21:17990.

2. CDC. Evaluation o a methodology or a collaborative multiple sourcesurveillance network or autism spectrum disordersAutism and

Developmental Disabilities Monitoring Network, 14 sites, UnitedStates, 2002. In: Surveillance Summaries, February 9, 2007. MMWR2007;56(No. SS-1):2940.

3. CDC. Prevalence o aut ism spectr um disorders Autism andDevelopmental Disabilities Monitoring Network, six sites, UnitedStates, 2000. In: Surveillance Summaries, February 9, 2007. MMWR2007;56(No. SS-1):111.

4. CDC. Prevalence o aut ism spectr um disorders Autism andDevelopmental Disabilities Monitoring Network, 14 sites, UnitedStates, 2002. In: Surveillance Summaries, February 9, 2007. MMWR2007;56(No. SS-1):1228.

5. Yeargin-Allsopp M, Rice C, Karapurkar , Doernberg N, Boyle C, Murphy CPrevalence o autism in a US metropolitan area. JAMA 2003;289:4955

6. CDC, National Center or Health Statistics. Estimates o the July 1, 2000July 1, 2007, United States resident population romthe vintage 2007 postcensal series by year, county, age, sex, race, andHispanic origin, prepared under a collaborative arrangement with the U.SCensus Bureau. Bethesda, MD: U.S. Department o Health and HumanServices, CDC, National Center or Health Statistics; 2007. Availablat: http://www.cdc.gov/nchs/nvss/bridged_race/data_documentation.htm#vintage2007. Accessed December 7, 2009.
http://www.cdc.gov/nchs/nvss/bridged_race/data_documentation.htm#vintage2007http://www.cdc.gov/nchs/nvss/bridged_race/data_documentation.htm#vintage2007http://www.cdc.gov/nchs/nvss/bridged_race/data_documentation.htm#vintage2007http://www.cdc.gov/nchs/nvss/bridged_race/data_documentation.htm#vintage2007


27/28


28/28

U.S. Government Printing Oce: 2009-523-019/41218 Region IV ISSN: 1546-0738

MMWR

TeMorbidity and Mortality Weekly Report (MMWR) Series is prepared by the Centers or Disease Control and Prevention (CDC) and is available reeo charge in electronic ormat. o receive an electronic copy each week, visitMMWRs ree subscription page athttp://www.cdc.gov/mmwr/mmwrsubscribe.html.Electronic copy also is available rom CDCs Internet server at http://www.cdc.gov/mmwror rom CDCs le transer protocol server attp://tp.cdc.gov/pub/

publications/mmwr. Paper copy subscriptions are available through the Superintendent o Documents, U.S. Government Printing Oce, Washington, DC20402; telephone 202-512-1800.

Data in the weeklyMMWRare provisional, based on weekly reports to CDC by state health departments. Te reporting week concludes at close o business

on Friday; compiled data on a national basis are ocially released to the public on the ollowing Friday. Data are compiled in the National Center or PublicHealth Inormatics, Division o Integrated Surveillance Systems and Services. Address all inquiries about theMMWRSeries, including material to be consideredor publication, to Editor,MMWRSeries, Mailstop E-90, CDC, 1600 Cliton Rd., N.E., Atlanta, GA 30333 or to [email protected].

All material in theMMWRSeries is in the public domain and may be used and reprinted without permission; citation as to source, however, is appreciated.

Use o trade names and commercial sources is or identication only and does not imply endorsement by the U.S. Department o Health and HumanServices.

Reerences to non-CDC sites on the Internet are provided as a service toMMWRreaders and do not constitute or imply endorsement o these organizationsor their programs by CDC or the U.S. Department o Health and Human Services. CDC is not responsible or the content o these sites. URL addresseslisted inMMWRwere current as o the date o publication.
http://%20http//www.cdc.gov/mmwr/mmwrsubscribe.htmlhttp://www.cdc.gov/mmwrftp://ftp.cdc.gov/pub/publications/mmwrftp://ftp.cdc.gov/pub/publications/mmwrmailto:[email protected]:[email protected]://ftp.cdc.gov/pub/publications/mmwrftp://ftp.cdc.gov/pub/publications/mmwrhttp://www.cdc.gov/mmwrhttp://%20http//www.cdc.gov/mmwr/mmwrsubscribe.html

CDC Report 57 Percent Increase in Autism

Documents

Transcript of CDC Report 57 Percent Increase in Autism