The Feasibility of Using Electronic Health Data for Research on Small Populations. Extent of coverage in current federal data collection activities


Existing data collection efforts have several foci and purposes, and their strategies reflect the challenges just mentioned. The Centers for Disease Control and Prevention collect data about the prevalence of ASDs through 14 sites in the Autism Developmental Disabilities Monitoring Network, which identify 8-year-old children with ASDs and other developmental disabilities through record review every other year. Most national health-related surveys do not have a longitudinal design, making it impossible to follow cohorts of youth with ASDs as they transition to adulthood. However, the Department of Education conducts several longitudinal studies, including as the National Longitudinal Transition Study (NLTS), the Early Childhood Longitudinal Survey (ECLS), and the National Household Education Surveys (NHES). These surveys focus on services that children and youth receive in school and the effects of childhood disability on adult outcomes.140 The National Longitudinal Study follows a national sample of students who were 13 to 16 years old in 2000. The Early Childhood Longitudinal Study includes 3 cohorts of children who were either followed from birth through kindergarten or from kindergarten through grade school. The survey asks ASD-related questions of the early childhood (9 months-kindergarten) cohort. The National Household Education Survey collects information from adults on learning at all ages among members of their household, from early childhood through school-age and adulthood, capturing a sample of adequate size for national and regional estimates.

The major sources of health-related data have been the National Longitudinal Survey of Adolescent Health (a longitudinal, cohort study that began in the 1994–1995 school year to follow a nationally representative sample of adolescents in grades 7–12 in the United States) and the National Survey of Children’s Health (NSCH, collected from a random sample of households) which focuses on physical limitations, symptoms, and diagnoses. The NSCH also collects information on medications prescribed, services used, and more general questions on health and health care.141 The State and Local Integrated Telephone Survey (SLAITS) is also an important source of information on children with special health needs, such as those with ASDs. It is collected to supplement national data by providing more detailed information from states (see Table I.3).

There is a survey question that asks parents if their child has ever been diagnosed with autism, Asperger’s disorder, pervasive developmental disorder, or other autism spectrum disorder, but only one yes or no response for the list of conditions is recorded. The validity of parental reports of such diagnostic information is also open to question and affects the official counts of ASD prevalence.142

Survey data collection about children and adolescents on the autism spectrum faces a distinctive set of challenges. First is the problem of small numbers, with ASDs, according to CDC estimates, occurring in only about one of every hundred households that might be contacted for a survey. Second, the condition can be difficult to diagnose (the state-of-the-art diagnostic regime takes several hours to administer) and diagnostic criteria are evolving. There are thus concerns about the validity and reliability of reported cases. The issue of missed diagnoses is a large problem, but may be decreasing as available services, support, and ASD awareness grow.143 Third, as is the case with many child health issues, data cannot be collected directly from the affected individuals because of their age or the nature of their disorders. Data must be collected from a proxy, generally a parent, school, or service provider. Clinicians typically identify children with ASDs when they fail to meet specific developmental progress milestones, or when certain behavioral characteristics are observed. Federal surveys identify children with ASDs by asking the parent, “Has a doctor or health professional ever told you that your child has autism?” The alternative sources of data may not apply the same diagnostic criteria. There may be tendencies to over-report because of eligibility for sought-after special programs and resources. Conversely, there may be under-reporting in demographic groups not aware of these services, as well as in the past when awareness around ASDs was less prevalent.144 Shifting definitions and lack of biologic markers for ASDs has made identification difficult,145 and getting a consistent definition applied across respondents is no small challenge.

There are additional challenges in identifying ASDs in racial and ethnic minorities because some observed characteristics may be attributed to cultural norms and communication barriers related to immigration status or ethnicity.146 These challenges may result in under diagnosis of ASDs in minority groups.147 One study found pediatricians were more likely to suspect autism when judging clinical vignettes of ethnic majority cases than among minorities.148

Obtaining data on youth with ASDs (or other disabilities) as they transition to adulthood faces additional complexities. The activities of daily living that may be used as measures of disability differ for youth and adults, and youth and adults are eligible for different programs and program participation may be used as an indicator of disability.149

View full report


"rpt_ehealthdata.pdf" (pdf, 1.99Mb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®