 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
a Children's Institute, Rochester, New York
b Department of Pediatrics, School of Medicine and Dentistry, University of Rochester, Rochester, New York
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
OBJECTIVE. The purpose of this work was to estimate the loss of household income associated with childhood autism using a nationally representative sample.
METHODS. Parents of 11684 children enrolled in kindergarten to eighth grade were surveyed by the National Household Education Survey-After School Programs and Activities in 2005. An autism spectrum disorder was defined as an affirmative response to the questions, "has a health professional told you that [child] has any of the following disabilities? 1) autism? 2) pervasive developmental disorder or PDD?" There were 131 children with autism spectrum disorder in the sample and 2775 children with other disabilities. We used ordinal logistic regression analyses to estimate the expected income of families of children with autism given their education level and demographic characteristics and compared the expected income with their reported income.
RESULTS. Both having a child with autism spectrum disorder and having a child with other disabilities were associated with decreased odds of living in a higher income household after controlling for parental education, type of family, parental age, location of the household, and minority ethnicity. The average loss of annual income associated with having a child with autism spectrum disorder was $6200 or 14% of their reported income.
CONCLUSION. Childhood autism is associated with a substantial loss of annual household income. This likely places a significant burden on families in the face of additional out-of-pocket expenditures.
Key Words: autism • household income • pervasive developmental disorder • national survey • economics • National Household Education Survey • United States
Abbreviations: ASD—autism spectrum disorder • PDD—pervasive developmental disorder • NCES—National Center for Education Statistics • NHES—National Household Education Survey • CI—confidence interval • OR—odds ratio
Autism spectrum disorders (ASDs) are pervasive developmental disorders that are characterized by limited verbal and nonverbal communication, social reciprocity, and restrictive and repetitive behaviors. ASD include autism, Asperger's syndrome, and pervasive developmental disorder (PDD). The impact of childhood autism on the familial economy is presumed to be large. On the expenditure side, the literature reports both expenditures paid by parents and by taxpayers. In 1999–2000, taxpayers in the United States spent $12773 on service expenditures associated with education for a child with autism.1 A pilot study based on a small sample of children with autism in the United Kingdom2 reported substantial out-of-pocket educational expenditures paid by parents. Out-of-pocket expenditures for educational and behavioral supports by parents of children with ASD have not been quantified in the United States; however private-pay tutoring, private schooling, speech or occupational therapy, and similar expensive services are commonly used. In addition to educational and behavioral services, families with children with autism spend more in health care services than other families. Using nationally representative samples from the Medical Expenditure Panel and the National Hospital Ambulatory Medical Care Survey, 1 recent study3 estimated that parents of children with ASD spend $5272 more in health care services than other families. Other studies4,5 have shown that health care expenditures are substantially higher for children with ASD, as well as for children with disabilities in general. We know far less about the impact of childhood autism on the income side of the familial economy. The British pilot study2 reported a median weekly loss of income of £250. There are no studies estimating loss of income associated with childhood autism in the United States.
Our objective in this study was to estimate the average loss of household income associated with childhood autism in the United States using a nationally representative sample. We used a proportional odds ordinal logistic model to estimate the loss of income attributable to having a child with ASD by modeling expected income values based on parental education and demographic characteristics of families of children with ASD and comparing results with the income these families reported.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
15 years of age. This survey was developed by the National Center for Education Statistics (NCES) to measure participation in after-school programming.6 It is part of the NHESs, which have been conducted biannually in the United States since 1991. The survey interviewed parents of children aged 5 to 15 years between January and April 2005. The respondent was the adult in the household most knowledgeable about the sampled child's care and education, typically the mother. The response rate for the After School Programs and Activities Survey was 84.1%.6 Weights were provided to adjust for the complex survey design and for nonresponse rates. Thus, the estimates presented in this article are nationally representative estimates.
Measures
ASD Measures
Parents were asked to respond to the questions, "has a health professional told you that [child] has any of the following disabilities? 1) autism? 2) pervasive developmental disorder or PDD?" We defined ASD as any affirmative response to these questions with parents indicating autism, PDD, or both autism and PDD.
Household INCOME
We measured household income using a 14-category ordinal question, "what was the total income of all persons in your household over the past year, including salaries or other earnings, interest, retirement, and so on for all household members?" Categories 1 to 10 classify the household income at $5000 intervals from 0 to $50000, category 11 has a $10000 spread (from $50001 to $60000), category 12 has a $15000 spread (from $60001 to $75000), category 13 has a $25000 spread (from $75001 to $100000), and category 14 is censored (over $100000). To calculate estimates of lost income, we calculated the midpoint of each category and conservatively set the last category midpoint at $101000, thus censoring the right tail of the income distribution and eliminating the influence of outliers. Median estimates of income loss were not used, because they are exogenously determined by the spread of the categories of the income question.
Other Disabilities
Parents were asked if their child had any of the following conditions: "a specific learning disability," "mental retardation," "speech or language delay," "serious emotional disturbance," "deafness or another hearing impairment," "blindness or another visual impairment," "an orthopedic impairment," "attention deficit disorder, ADD or [attention-deficit/hyperactivity disorder] ADHD," and "other health impairment lasting 6 months or more." Children whose parents reported to have any of these conditions and did not have ASD were classified in the other disability group.
Demographic Information
The average age of parents was calculated by the mean of the maternal and paternal ages for 2-parent households and the age of the mother or father for single-parent households. Type of family was measured with a dichotomous variable separating 2-parent families from all of the other family types.
NCES provided a derived variable that, "indicates the highest level of education for the subject child's parents or nonparent guardian who reside in the household."7 The variable has 5 categories: (1) less than a high school diploma, (2) high school graduate or equivalent, (3) vocational or technical college or some college education, (4) college graduate, and (5) graduate or professional school.
We used the race of the child as a proxy variable for the minority status of the parents, because information on race of the parent was not collected in the survey. We dichotomized the race variable to reflect minority ethnicity, which included black, Asian or Pacific Islander, all other nonwhite races, multiple races, and all Hispanic children of any race.
All of the participants provided informed consent for the survey. Additional information on the survey and informed consent procedures is available.6
Weights
We used Stata (Stata Corp, College Station, TX) to adjust for the complex sample design using Taylor approximations that provide the correct standard errors, following NCES guidelines.6
Econometric Analysis
The econometric analysis required several steps. First, we tested the proportional odds assumption. We then estimated the proportional odds ordinal logistic model on household income using average parent age, type of family, race (based on the race of the child), level of parent education, urban or rural living area, ASD status, and other disability status as independent variables. The ordinal logistic model is appropriate because household income was measured with a 14-category ordinal variable with unequal spreads (10 categories at $5000 and 5 categories higher than $5000). The model estimated 13 simultaneous logistic regressions imposing the restriction that odds ratios for independent variables be the same (proportional odds assumption). The model was then used to predict the probability of falling into each of the 14 income categories for any value of the independent variables. Using the income category midpoints and the estimated probabilities for each category, we calculated average household income based on a censored income measure at $101000.
Next, we validated the model by estimating the average censored income for the average person in the sample (all independent variables at their average level) and comparing it with the average reported income in the sample. We calculated the average values of the independent variables for households with a child with ASD and other disability. Using these values, we estimated the expected income of these households given their education level and demographic characteristics absent ASD or disability. We then compared these estimates of expected income with their reported income. The difference between estimated expected values and reported values was the loss of income associated with having a child with ASD or other disability.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Median Household Income
The median income for the sample was $47500, and the mean censored income was $53254. The median income for the US households in 2005 was $46326 according to the US Census Bureau.8
Household Characteristics
Table 1 describes the demographic characteristics of the sample for nondisabled children, children with other disabilities, and children with ASD. There were no statistically significant differences between the ASD group and any of the other groups, likely in part related to the relatively small sample of children with ASD. However, more children with other disabilities lived in a household with lower parental education level (28.7% high school graduate vs 24.6%) and lived in an urban area (23.8% vs 20.0%) compared with children without disabilities. Children with other disabilities were less likely to live in 2-parent households (68.2% vs 73.3%) than children without disabilities. We found no differences by race. There also were no significant differences by average parent age among the 3 groups of households.
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Economic theory identifies 3 sources of lower-than-expected household income: (1) poorer-than-expected labor market performance, (2) lower-than-expected labor participation, and (3) lower-than-expected savings and investment, based on parent demographic characteristics and level of education. All 3 of these issues may be true for parents of children with ASD.
Recent research has shown that parents of children with emotional and behavioral disorders lack appropriate community-based services and resources needed to support work and family obligations.9 Families with a child with a serious disability often accommodate family and work obligations to the requirements and behavior of the affected child.10 Lack of resources that fit the special needs of the child can have a significant impact on work and family functioning, leading to significant difficulty in establishing a work-family balance.11 We suspect that this is true for many families of children with autism. For example, in a nationally representative study,12 fathers of children with autism were less likely to report full-time employment compared with fathers of the nonautistic population and were more likely to work part time. Thus, we speculate that the most likely explanation for loss of annual income in the context of childhood autism is that parents of a child with ASD make different working choices than other parents,12 perhaps because of the needs associated with ASD combined with the lack of appropriate community-based services and resources.
We also speculate that having a child with ASD likely results in lower savings and investment by families, because of the reported higher medical, educational, and behavioral expenses, resulting in lower interest and dividend income in future periods. There are no studies on the saving behavior of households with a child with ASD or studies of the labor market performance of parents of children with ASD; thus, these speculations merit additional research.
There are alternative explanations for our findings. The association between childhood autism and household income may not represent an indirect cost related to burden of care of a child with ASD but could be the result of other causes. In particular, parental variables related to the broader autism phenotype could be directly associated with unemployment and lower income.12,13 In addition, some families with children with ASD may be strategically earning less income to remain eligible for health or other autism treatment-related benefits. Although we consider both of these explanations less likely than the burden of care explanation, they all merit scientific study.
Regardless of the reason for loss of income, however, our findings suggest a significant burden for families.
Limitations of the Study
There are a few potential limitations in this study. Although parent report of ASD is viewed as a fairly reliable,14,15 we did not have access to medical charts or diagnostic reports, and we were limited by the way the ASD and income questions were asked in the survey. In addition, we had a relatively small sample of children with ASD, and, thus, could not investigate additional differences by subgroup. Although the NHES did not ask about Asperger's syndrome, the high prevalence of ASD in the sample suggests that children diagnosed with Asperger's syndrome were included in the autism measure. In addition, we relied on parent report of income, and reporting bias may have been present. Higher income households may be less likely to report a child with disabilities. However, we have no reason to suspect that reporting accuracy would differ for parents of children with autism as compared with other parents.
Implications
Childhood ASDs are associated with a substantial loss of household income. This suggests the need for additional evaluation of available supports for families, as well as specific barriers to optimizing family income. In the meantime, it is prudent for health care providers to ask families of children with autism about financial difficulties and to assist them in obtaining access to appropriate health care, educational services, and community resources.
|
FOOTNOTES |
|---|
Address correspondence to Guillermo Montes, PhD, Children's Institute, 271 N Goodman St, Suite D103, Rochester, NY 14607. E-mail: gmontes{at}childrensinstitute.net
The authors have indicated they have no financial relationships relevant to this article to disclose.
| What's Known on This Subject Families of children diagnosed with an autism spectrum disorder spend $5272 more in health care services than other families. In addition, families of children with autism spectrum disorder face higher expenses for behavioral and educational treatments than other families.
|
| What This Study Adds Childhood autism is associated with a loss of annual household income of $6200 or 14% of their reported income. This places a significant burden on families who face additional out-of-pocket expenses with smaller incomes.
|
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2008 American Academy of Pediatrics. All rights reserved. |
||
|
||
