OBJECTIVE: The number of individuals diagnosed with autism spectrum disorders (ASDs) continues to increase in the United States and other developed countries; however, ASD is diagnosed less commonly in Hispanic than in non-Hispanic white individuals. This report analyzes differences in ASD prevalence between Hispanic and non-Hispanic whites in a large, population-based sample of 8-year-old children, and explores how prevalence has changed over time.
METHODS: Population-based surveillance of ASD was conducted on 142 717 8-year-old children. Evaluation of clinical and educational records resulted in 1212 children meeting the case definition criteria in 4 study years between 2000 and 2006.
RESULTS: ASD prevalence in Hispanic children was lower than in non-Hispanic white children (P < .005) for all study years. More Hispanic than non-Hispanic white children met the case definition for intellectual disability (P < .05) in study years 2004 and 2006. Prevalence of ASD diagnosis increased in both groups; the Hispanic prevalence almost tripled, from 2.7 per 1000 in 2000 to 7.9 per 1000 in 2006. A comparison of prevalence ratios found that Hispanic and non-Hispanic white ASD prevalence became significantly more similar from 2000 to 2006 (χ2 = 124.89, P < .001).
CONCLUSIONS: The ASD prevalence for Hispanic individuals in this population-based sample is substantially higher than previously reported. Nonetheless, Hispanic children continue to have a significantly lower ASD prevalence in comparison with non-Hispanic whites. The prevalence of ASD is increasing in both populations, and results indicate that the gap in prevalence between groups is decreasing.
- ADDM —
- Autism and Developmental Disabilities Monitoring Network
- ADDSP —
- Arizona Developmental Disabilities Surveillance Program
- ASD —
- autism spectrum disorder
- CI —
- confidence interval
- ID —
- intellectual disability
- NH white —
- non-Hispanic white
- OR —
- odds ratio
What’s Known on This Subject:
The number of individuals diagnosed with autism spectrum disorders (ASDs) continues to increase in the United States and other developed countries. Most prevalence estimates indicate that ASD is diagnosed less commonly in Hispanic individuals compared with non-Hispanic (NH) white populations.
What This Study Adds:
Prevalence of ASD in Arizona’s population-based cohort is higher than reported previously. Prevalence in the Hispanic population and NH white population increased significantly over time, with a significant decrease in the gap between Hispanic and NH white prevalence.
Autism spectrum disorders (ASDs) are a group of neurodevelopmental disorders characterized by abnormalities in communication, social impairments, and unusual or stereotypical behaviors. ASDs have been the subject of increasing scrutiny over the past decade, because the number of children identified as having an ASD has risen dramatically.1–9 However, recent studies of ASD prevalence have shown significant discrepancies between the prevalence of ASD in different ethnic and racial groups.1–3 In particular, ASD prevalence in Hispanic populations appears lower than ASD prevalence in the non-Hispanic white population.10–13 This difference may reflect a population difference in ASD prevalence between non-Hispanic white (NH white) and Hispanic groups as a result of genetic and environmental risk factors. Alternatively, it may be explained by socioeconomic issues, differing health behaviors, or other unknown factors that impede our ability to identify ASD prevalence accurately.
The Arizona Developmental Disabilities Surveillance Program (ADDSP) is a population-based surveillance system for ASD and intellectual disabilities that utilizes data from 15 school districts and clinical sources in Maricopa County, Arizona, the most populous county in Arizona, with ∼100 000 annual births. ADDSP activities are conducted in conjunction with the Autism and Developmental Disabilities Monitoring Network (ADDM), a multisite network funded by the Centers for Disease Control and Prevention that uses a standardized protocol for ASD identification.14 Prevalence estimates of the ADDM network have been published for children age 8 in 2000, 2002, 2004, and 2006.1–3 The purpose of this report is to use data from the ADDSP to investigate differences in ASD prevalence between Hispanic and NH white 8-year-old children and to explore time trends in prevalence between these ethnic groups. These data have implications for identification of ASD in clinical and educational settings, and for future investigations of modifiable risk factors for ASD in both NH white and Hispanic populations.
The ADDM surveillance system ascertains ASD among 8-year-old children through systematic review of clinical and educational records. Methods are modeled after the Centers for Disease Control and Prevention’s Metropolitan Atlanta Developmental Disabilities Surveillance Program, and have been described in detail.14 In brief, the ADDM protocol involves clinical and educational record abstraction and expert clinician review. Educational records of children who are 8 years old in the appropriate study year (2000, 2002, 2004, or 2006) and receive special educational services are reviewed. Clinical records of all children born in the appropriate study year, who received clinical services before the end of that study year, and whose record cites 1 or more of an extensive list of International Classification of Diseases, Ninth Revision diagnostic codes are also reviewed. Educational or clinical records that document a previous ASD diagnosis, results of tests commonly used to diagnose or evaluate ASD, or any of 34 key behaviors typical for ASD are abstracted. Data from all sources are deidentified and examined by a specially trained clinician reviewer to determine whether the child meets ASD case definition. Abstraction of variables related to intellectual disability (ID) was also performed for all children whose cases were abstracted for ASD, although clinician review for ID did not occur in 2000. The surveillance population for the 2004 study year was smaller than the other study years and was completed by only a subset of ADDM states with sufficient resources, including Arizona.
Coding of Race and Ethnicity
Race and ethnicity were coded during abstraction of records at school or clinical sources. In a minor modification of guidelines by the Office of Management and Budget, children were classified as white, non-Hispanic; black, non-Hispanic; Hispanic, any race; American Indian or Alaska Native, non-Hispanic; Asian or Pacific Islander, non-Hispanic; and Other or mixed race, non-Hispanic.15 Child race and ethnicity were coded according to official documents in the child’s record (such as an Individualized Education Plan or intake forms). If official sources within the record noted discrepant races, but each source noted the ethnicity as Hispanic, the child was coded as Hispanic. However, if sources within the record did not agree as to the race, and the child was not identified as Hispanic, the child’s race and ethnicity was coded as “discrepant”. Where available, Arizona birth certificate information was obtained in cases of missing or discrepant race or ethnicity in the child’s record. If either parent was listed as Hispanic on the birth certificate, the child was coded as Hispanic. A child was coded as white only if both parents were listed as white and non-Hispanic.
Data Analysis Plan
Census data were used to estimate the total number of 8-year-old children in the study catchment area in each study year; this number was used as the denominator to calculate the prevalence of ASD. The numerator included all children who met the ASD case definition based on clinician review. Frequency counts and χ2 comparisons of groups were calculated by using SPSS version 18.0. The ratio of NH white to Hispanic children with an ASD was assessed by using odds ratios (OR) obtained with Epi Info version 3.5.1. Epi Info was also used to evaluate the trends in prevalence over time by using the χ2 test for trend over time and the Mantel-Haenszel χ2 test,16 both tests that can be used to evaluate a linear increase in prevalence.6 ASD prevalence for the NH white and Hispanic populations was compared for the whole population, as well as separately for boys and girls, and changes in prevalence ratios were examined by using the Mantel-Haenszel χ2 test.
As seen in Table 1, the total number of 8-year-old children residing in the study area for all study years (2000, 2002, 2004, 2006) was 142 717. Of those children, 75 545 (52.9%) were NH white children, and 52 753 (37.0%) were Hispanic children. The study area population has a higher percentage of Hispanics than the overall population of Arizona: in the 2000 census, NH whites made up 63.8% of the population and Hispanics made up 25.2%, with a smaller representation of other ethnic groups such as African Americans and Asian Americans.17
Overall, 2.4% of cases had missing or discrepant race information. Because these cases represented a small percentage of total ASD cases, their impact on prevalence estimates was negligible. Among children living in the catchment area from 2000 to 2006, 1212 met ASD case definition as determined by the ADDM protocol.
Characteristics of the Hispanic and NH white children who met ASD case definition are presented in Table 2 by study year. Among children who met criteria for ASD in all study years, both Hispanic and NH white samples showed a significantly greater number of boys than girls, with a boy-to-girl ratio among Hispanics ranging from 1.79 to 2.94, and a ratio among NH whites ranging from 3.80 to 5.72. Across all study years the boy-to-girl ratio was more pronounced in the NH white population than in the Hispanic population. Among children who met the ASD case definition and whose records were also reviewed to ascertain case definition for ID (a subset of the total ASD population), a similar proportion of Hispanic and NH white children met the case definition for a co-occurring ASD and ID in 2002, but, a significantly higher proportion of Hispanic children met the case definition for a co-occurring ASD and ID in 2004 and 2006. Across all study years a significantly higher percentage of families of Hispanic children with ASD had incomes below the 2000 national average of $50 000 per year.
ASD Prevalence Trends
ASD prevalence for each study year is presented in Table 3 and Fig 1. Hispanic ASD prevalence, ranging from 2.7 to 7.9 per 1000, was higher than has been reported in previous investigations.10–13 For all study years, prevalence in Hispanic children was lower than in NH white children (P < .005). χ2 tests of trend over time indicate that ASD prevalence in both Hispanic (χ2 = 50.98, P < .001) and NH white (χ2 = 52.37, P < .001) samples increased significantly between 2000 and 2006.
Hispanic and NH white samples were further analyzed by gender. In each study year, prevalence of ASD for Hispanic boys was significantly lower than for NH white boys (Table 3). In contrast, among girls, Hispanic prevalence of ASD was significantly lower than among NH white girls only in study year 2000; in later years, prevalence of girls did not differ by ethnicity (Table 3). χ2 tests of trend over time indicate that in each gender and ethnicity group, prevalence increased significantly over the 6-year study period.
Comparison of Prevalence Ratios
The ratios of prevalence for ASD in NH white compared with Hispanic children were calculated for each study year. When both genders were included, Mantel-Haenszel χ2 tests of trend indicated that the OR of NH white to Hispanic children decreased significantly (χ2 = 124.89, P < .001) from 3.22 (95% CI = 2.26–4.61) in 2000 to 1.92 (95% CI = 1.56–2.37) in 2006. When analyzed by gender, the OR of NH white to Hispanic children with ASD also significantly decreased in both boys and girls. Among girls (χ2 = 4.95, P < .05), the NH white to Hispanic ratio decreased from 1.85 in 2000 (95% confidence interval [CI] = 0.98–3.53) to 1.37 in 2006 (95% CI = 0.88–2.15). The ratio of NH white to Hispanic boys also showed a significant decrease over time (χ2 = 128.75, P < .001), dropping from 4.04 in 2000 (95% CI = 2.60–6.30) to 2.11 in 2006 (95% CI = 1.66–2.68). Regression lines used to test for a linear decrease in prevalence ratios between 2000 and 2006 found a nonsignificant decrease.
The prevalence of ASD in this population-based cohort showed significant increases in 4 study years between 2000 and 2006, which is consistent with similar reports from across the United States and other developed countries.1–9 The ASD prevalence for Hispanics in the current study, however, is markedly higher than has been reported previously.10–12
The increased prevalence in Hispanic individuals that we observed in comparison with previous reports may be explained in part by different methods for estimating prevalence. Examining telephone survey results from the National Survey of Children’s Health between 2003 and 2004, Liptak and colleagues10 found that 2.6 per 1000 Latino children had a diagnosis of autism, per caregiver report.10 Though this result is lower than Hispanic prevalence noted in the current study, at least some of this difference may result from Liptak and colleagues’10 focus on autism rather than on the current study’s broader case definition of ASD. In addition, Liptak and colleagues10 relied on caregiver response, whereas the current study examined clinical review of case records from schools and clinical sources, which sometimes identified a child with ASD who had not previously been diagnosed. In a similar investigation using data from the National Health Interview Survey, the Hispanic ASD prevalence between 1997 and 2008 was estimated as 3.2 per 1000 children,18 similar to prevalence estimates from the current study in 2000 and 2002, but lower than the prevalence in 2004 and 2006. Low ASD prevalence among Hispanic populations has also been reported outside the United States; Montiel-Nava and Peña identified ASD in 1.7 per 1000 children in Maracaibo county, Venezuela.12 Although they ascertained ASD cases in a manner similar to the current study, some aspects of the Venezuelan health care system may have resulted in limited access to all ASD cases.12
Despite the higher prevalence of ASD in the current study, Hispanic children continue to show significantly lower prevalence in comparison with NH whites.10–13 The reasons for these differences are likely complex and multifaceted. Some recent investigations challenge the notion that Hispanic populations have a lower prevalence of ASD and suggest that reported rates are falsely low as a result of small sample sizes or limited access of the Hispanic community to appropriate health care facilities where ASD may be diagnosed.19 Many of these disparities may be economically driven. Other disparities in health care access and utilization between Hispanic and NH white populations may also contribute to these differences. For example, minority individuals are more likely to receive care in underresourced hospitals,20 and doctors refer Hispanic children less frequently to specialists than their NH White peers.21 Hispanic individuals are also less likely to be medically insured than NH whites,21 and non–English-speaking individuals may experience multiple barriers to medical care.22,23 Although it was not the focus of the current study, more research is warranted examining the potential association between health care disparities and ASD prevalence.
Previous investigations of adverse health outcomes in Hispanic populations have found a lower than expected prevalence of such outcomes when compared with other ethnic groups, despite more risk factors such as lower socioeconomic status, reduced access to medical care, and fewer years of formal education.24,25 For example, in comparison with NH whites, Hispanic populations have lower mortality rates,26,27 lower incidence of mental health disorders,28 and a lower risk of several types of cancer.29 Despite the possibility that these differences may result from a decreased susceptibility to specific diseases, any genetic factors that might explain them are largely unknown. It is also important to note that the Hispanic population represents a heterogeneous group of individuals from a multitude of countries, cultures, and historical backgrounds, which could contribute to genetic diversity among populations labeled as Hispanic. Lower ASD prevalence among Hispanic populations is also consistent with an observed tendency for increased risk of ASD among individuals with higher family income and more years of formal education.30–32 The origins of this trend are not well understood, but may relate to increased access to diagnostic services, or increased familiarity with and acceptance of an ASD diagnosis among those who are more educated and affluent.
The results of the current study corroborate other reports of increasing ASD prevalence in multiple ethnic and racial groups.4,31 In contrast to the highly significant differences in ASD prevalence in every study year for affected boys of different ethnicities, the difference among affected girls was significant for only 1 study year. As a result of these differences, the boy-to-girl ratio was substantially higher for NH white children with ASD in comparison with Hispanic children. Although the boy-to-girl ASD prevalence ratio is consistently reported to be 3:1 or greater,1–3 there is little research about the interaction of gender and ethnicity differences. The data herein reported suggest that the driver of these differences is a persistently higher prevalence in NH whites compared with Hispanic boys and that additional investigations into this phenomenon are warranted.
These results also suggest that the difference between Hispanic and NH white prevalence is decreasing. The decrease in the OR between NH white and Hispanic ASD prevalence appears largely due to the marked increase in Hispanic prevalence between 2000 and 2006. Although the reasons for increased ASD prevalence are unknown for all populations, it is possible that the increased prevalence in the Hispanic population may result from increased awareness of ASD among parents and health care professionals who work with Hispanic children, improved access to ASD-related services in the Hispanic population, and other related medical and educational factors. To explore this possibility further, we analyzed how the percentage of children who had a previous diagnosis of ASD (before they were identified by the ADDSP) differed over time by ethnicity. Between study years 2000 and 2006, rates of previous ASD diagnosis in the Hispanic population increased from 38% to 49% of children who met the case definition for ASD, and rates of previous diagnosis in the NH white population increased from 53% to 57%. These results suggest that the increase in Hispanic prevalence may be explained at least in part by increased awareness of ASD and an associated increase in community diagnosis.
Intellectual Disability and ASD
Findings from the current study indicate that more Hispanic than NH white children met ADDM criteria for both ID and ASD in 2004 and 2006. A number of interesting characteristics have been identified in individuals with co-occurring ID and ASD.31,33,34 For example, people with both ID and ASD are reported to have significantly poorer social interactions than their peers with ID only and may also show more restricted interests and repetitive behavior, although both groups often have similar communication.33–35 In keeping with the results of the current study, McInnes and colleagues36 found that 21 of 35 (60%) children with ASD in a Costa Rican sample had an IQ <50. This is consistent with other investigations of ID and ASD co-occurrence, in which children of Hispanic origin were at increased risk for severe ID in comparison with NH white children.37 However, Mandell and colleagues,11 by using data from the ADDM network, found that among children with IQs <70, Hispanic children were less likely than NH white children to have a previous diagnosis of ASD. More surveillance research on co-occurring ASD and ID is necessary to improve our understanding of prevalence and characteristics as they relate to race and ethnicity.
Strengths and Limitations
One of the primary strengths of this study is its population-based data collection, which eliminates the referral bias of most previous investigations. In addition, the systematic and externally validated nature of the case finding and case identification procedures minimizes the clinical variability of referral samples. A potential limitation of this study is that individual cases are not validated prospectively by clinician evaluation. A recent analysis of the case definition protocol used in this study, however, shows that the ADDM protocol likely underestimates the number of children diagnosed with ASD, in comparison with diagnosis by an experienced clinician who is blinded to case definition status.38 Another potential limitation of this investigation is that the results may not be generalizable to all populations of Hispanic origin. The Hispanic population of Arizona primarily comprises individuals of Mexican descent. It remains to be seen whether a similar prevalence will be found in Hispanic populations whose members come from other countries of origin and may have different sociodemographic characteristics. Finally, as a result of our data acquisition methods, the findings of the study are necessarily limited to the accuracy of the information available within children’s records.
The results of this study have important implications for future ASD research and clinical practice. Additional investigation of the underlying genetic, environmental, and systems-based influences on the ethnic disparities in ASD prevalence has the potential to identify modifiable risk factors that may improve clinical outcomes. Clinically, the results of the current study underscore the importance of continuing to improve access to quality health care for people across all racial and ethnic populations, in particular, given the increasing prevalence of ASD in the Hispanic population. For example, surveillance for ASD would likely benefit from improved screening measures available in both English and Spanish. Also, the high rate of co-occurring ID and ASD in the Hispanic population indicates unique educational and health care needs that must be met by qualified professionals.
The prevalence of ASD in the Hispanic population appears to be higher than previously reported, even though Hispanic children have lower ASD prevalence when compared with non-Hispanic Whites. Building on previous reports that examined single point-in-time prevalence, results of the current study show that ASD prevalence in non-Hispanic white and Hispanic populations is rising, and that the prevalence gap between Hispanic and non-Hispanic white individuals is decreasing.
The authors thank Zhenqiang James Lu for his expert statistical consultation.
- Accepted November 17, 2011.
- Address correspondence to Christopher Cunniff, MD, Department of Pediatrics, University of Arizona, 1501 N. Cambell Ave, PO Box 245073, Tucson AZ 85724. E-mail:
Dr Pedersen made substantial contributions to conception and design, analysis and interpretation of data, and drafting the article, and approved the final version to be published; Drs Meaney and Cunniff made substantial contributions to conception and design, analysis and interpretation of data, acquisition of data, and drafting the article, and approved the final version to be published; Dr Pettygrove made substantial contributions to conception and design, analysis and interpretation of data, acquisition of data, critical revision of the manuscript for important intellectual content, and approved the final version to be published; Ms Mancilla and Gotschall made substantial contributions to acquisition of data, analysis and interpretation of data, critical revision of the manuscript for important intellectual content, and approved the final version to be published; Drs Kessler and Grebe made substantial contributions to acquisition of data, critical revision of the manuscript for important intellectual content, and approved the final version to be published.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by the Centers for Disease Control and Prevention.
- Autism and Developmental Disabilities Monitoring Network 2006 Principal Investigators.
- Autism and Developmental Disabilities Monitoring Network Surveillance Year 2000 Principal Investigators.
- Autism and Developmental Disabilities Monitoring Network Surveillance Year 2002 Principal Investigators.
- Montiel-Nava C,
- Peña JA
- Office of Management and Budget
- Szlko M,
- Nieto J
- ↵US Census Bureau. American FactFinder. 2011. Available at: http://factfinder.census.gov/home/saff/main.html. Accessed March 14, 2011
- Boyle CA,
- Boulet S,
- Schieve LA,
- et al
- Fombonne E
- Flores G,
- Abreu M,
- Tomany-Korman SC
- ↵Flores G, Tomany-Korman SC. The language spoken at home and disparities in medical and dental health, access to care, and use of services in US children. Pediatrics. 2008;121(6). Available at: www.pediatrics.org/cgi/content/full/121/6/e1703
- Blumberg SJ,
- Read D,
- Avila RM,
- Bethell CD
- ↵US Census Bureau. The 2012 Statistical Abstract. 2011. Available at: www.census.gov/compendia/statab/. Accessed November 10, 2011
- Arias E,
- Anderson RN,
- Hsiang-Ching K,
- Murphy SL,
- Kochanek KD
- Grant BF,
- Stinson FS,
- Hasin DS,
- Dawson DA,
- Chou SP,
- Anderson K
- Copyright © 2012 by the American Academy of Pediatrics