Title: Socioeconomic disparities in the age of first diagnosis of autism spectrum disorder (ASD) in Metropolitan Atlanta
1Socioeconomic disparities in the age of first
diagnosis of autism spectrum disorder (ASD) in
Metropolitan Atlanta
- Sally M. Brocksen, PhD
- Kimberly Kimiko Powell, PhD, RD
The findings and conclusions in this
presentation are those of the presenter and do
not represent those of the Centers for Disease
Control and Prevention
2Background and Purpose
- For children with an autism spectrum disorder
(ASD), early identification is crucial in
providing better developmental outcomes. - The impact of socioeconomic and demographic
factors needs to be examined to determine if an
ASD diagnosis is delayed within certain
populations postponing treatment and
intervention services. - This study examines whether there are differences
among children meeting the Metropolitan Atlanta
Developmental Disabilities Program (MADDSP)
surveillance case definition of ASD.
3Study design
- Children identified in the 2000 MADDSP study year
as having an ASD were linked with the 2000 census
data to analyze SES and demographic factors. - Block group census data on income, education,
occupation, employment, poverty status and
residential stability were analyzed using
principal component analysis (PCA) to create a
SES variable. - Additional regression analyses on demographic
factors (e.g. race, gender) were conducted to
evaluate differences among children identified as
having an ASD.
4MADDSP
- Design
- Ongoing, population-based, active monitoring
program based on record review - Mental retardation, cerebral palsy, vision
impairment and hearing loss autism spectrum
disorders since 1996 - Children aged 3-10 years, 1991-1994 8 year olds
in 1996, 2000, 2002 and future study years - Multiple sources (educational, clinical, service)
- Five counties in metro Atlanta
5MADDSP ASD Clinician Review Process
- Case status determined by systematic review of
abstracted information by autism/DD clinicians. - Behavioral coding scheme developed based on
DSM-IV, TR (2000) criteria for Autistic Disorder
and PDD-NOS. - All evaluation records for a child were compiled
and behaviors scored individually. - Criteria were summarized across evaluations to
determine case status. - Questionable cases are re-reviewed.
6Methods
- Children who meet the case definition of ASD may
or may not have a previous clinical diagnosis. - Children with this previous clinical diagnosis
are compared with children who have not received
a clinical diagnosis but meet the MADDSP case
definition of having an ASD.
7Methods
- Since individual level economic data was not
available this study used area-based measurements
from the 2000 census data to create a community
socioeconomic index. - A validated method used by Krieger (1992) was
employed to create socioeconomic profiles of the
neighborhoods (block groups) in which individuals
live.
8Methods
- Principal Component analysis (PCA) was used to
rank communities from low to high SES and
classified into tertiles using the following
variables - occupational class
- percent working class, professional class and
unemployed - income
- percent low income and percent high income
- poverty
- percent below poverty line
- education
- percent low and high education
- stability
- percent movement of houses and counties over five
years
9Results
N
Total number of children meeting the case definition of ASD 285
Children with a previous ASD diagnosis on a clinical evaluation 115 40.4
Children without a previous ASD diagnosis 170 59.6
ASD with IQ gt70 161 31.1
ASD IQ lt 70 107 20.7
Diagnosis before the age of five 56 19.6
10Results
Previous ASD diagnosis Previous ASD diagnosis No previous ASD diagnosis No previous ASD diagnosis ASD with IQ gt 70 ASD with IQ gt 70 ASD with IQ lt 70 ASD with IQ lt 70 Diagnosis before the age of five Diagnosis before the age of five
n n n n n
Sex
Male 104 26.3 137 43.2 144 35.7 82 20.3 50 48.1
Female 11 9.6 33 25.0 17 14.9 25 21.9 6 51.9
Race
White (non Hispanic 61 42.7 82 57.3 97 48.5 38 19.0 26 42.6
Black (non Hispanic) 38 37.3 64 62.7 41 20.1 57 27.9 22 57.9
Other 11 42.3 15 57.7 14 31.8 10 22.7 5 60.0
11Results
SES Tertile 1 1 2 2 3 3
n n n Total p-value
Total ASD cases 115 40.8 96 34.0 71 25.2 282 .001
Children with a previous ASD diagnosis 52 45.6 38 33.3 24 21.1 114 .299
Children without a previous ASD diagnosis 63 37.5 58 34.5 47 28.0 168 .299
ASD and IQgt70 79 49.4 52 32.5 29 18.1 160 .000
ASD and IQ lt 70 32 30.5 33 31.4 40 38.1 105 .204
Diagnosis before the age of five 26 47.3 21 39.2 8 14.5 55 .229
12Results
No previous ASD diagnosis No previous ASD diagnosis No ASD diagnosis before the age of 5 No ASD diagnosis before the age of 5 ASD with IQ gt70 ASD with IQ gt70
AOR p-value AOR p-value AOR p-value
Race, Black (non Hispanic) .92 (.49-1.74) .799 .31 (.11-.86) .025 .38 (.22-.65) .000
Race, Other .87 (.36-2.11) .762 .65 (.16-2.68) .553 .64 (.31-1.33) .234
Highest SES tertile .66 (.36-1.21) .182 1.05 (.41-2.65) .923 1.7 (.64-1.78) .801
Lowest SES tertile 1.21 (.61-2.41) .586 4.18 (1.26-13.9) .020 .60 (.34-1.08) .807
Male .40 (.19-.85) .018 1.57 (.40-6.25) .519 2.7 (1.53-4.90) .001
13Summary
- A child having ASD with an IQ gt 70 is associated
with being in the highest SES tertile whereas a
child having ASD with an IQ lt 70 is associated
with being in the lowest SES tertile. - Children in the lowest SES tertile are 4 times
more likely to have not received an ASD diagnosis
before the age of 5. - Findings are similar to the results by Karapurkar
Bhasin Schendel (in press) who examined
sociodemographic risk factors using 1996 MADDSP
study year data.
14Future studies
- Conduct analysis on future MADDSP study years to
look for trends in the diagnosing of ASD in
children from different socioeconomic groups. - Analyze the impact of specific census variables
related to SES (e.g. median household income). - Link data to birth certificate files to gain
information related to maternal age and
education. - Look at differences based on where the childs
record was obtained (school sources vs. clinical
sources).
15Strengths and limitations
- Strengths
- Used population based data to classify children
as ASD. - The creation of socioeconomic status variables
using multiple measurements. - Limitations
- No individual level economic data is available,
therefore census data was used to approximate the
socioeconomic status of identified children.
16Public Health Implications
- Early identification of ASD is crucial to
promoting optimal developmental outcomes for
children with ASD. - Strategies need to be developed to assess and
target the needs of children from all
socioeconomic strata.
17Acknowledgements
- Andrew Autry
- Jon Baio
- Claudia M. Bryant
- Matt Cahill
- Afiya Celestine
- Nancy Doernberg
- Shryl Epps
- Lekeisha Jones
- Rita Lance
- Charmaine McKenzie
- Catherine Rice
- Fiona Steele
- Darlene Sowemimo
- Melody Stevens
- Melissa Talley
- Ignae Thomas
- Kim Van Naarden-Braun
- Anita Washington
- Victoria Washington
- Laquita Williams
- Susan Williams
- Marshalyn Yeargin-Allsopp