BACKGROUND. The impact of childhood autism on parental employment is largely unknown.
OBJECTIVE. The purpose of this work was to describe the child care arrangements of children with autism and to determine whether families of preschool-aged children with autism are more likely to report that child care arrangements affected employment compared with typically developing children and children at high risk for developmental problems.
METHODS. Parents of 16282 preschool-aged children were surveyed by the National Survey of Children's Health. An autism spectrum disorder was defined as an affirmative response to the question, “Has a doctor or health professional ever told you that [child] has any of the following conditions? Autism?” There were 82 children with autism spectrum disorder in the sample, and 1955 children at high risk on the basis of the Parent's Evaluation of Developmental Status. We used χ2 and multivariate logistic regression analyses.
RESULTS. Ninety-seven percent of preschool-aged children diagnosed with autism spectrum disorder were cared for in community settings, particularly preschool and Head Start, with only 3% in exclusive parental care. Thirty-nine percent of the parents of children with autism spectrum disorder reported that child care problems had greatly affected their employment decisions, compared with 16% of the children at high risk and 9% of those who were typically developing. In multivariate analyses, families with a child with autism spectrum disorder were 7 times more likely to state that child care problems affected employment than other families, after controlling for household and child covariates. This effect was 3 times larger than the effect of poverty.
CONCLUSIONS. Developmental problems and autism spectrum disorder are associated with higher use of child care services and higher probability that child care problems will greatly affect employment. These findings warrant evaluation of the community resources available to families with children with special needs.
- child care
- pervasive developmental disorder
- national survey
- child care
- National Survey of Children's Health
- United States
Autism spectrum disorders (ASDs) are pervasive developmental disorders that are characterized by limited verbal and nonverbal communication, social reciprocity, and restrictive and repetitive behaviors. ASDs include autism, Asperger syndrome, and pervasive developmental disorder (PDD).
The impact of childhood autism on parental employment is largely unknown. A nationally representative study estimated that households with a child with autism earn 14% less1 income than demographically and educationally comparable households. In another population-based study, fathers of children with autism were less likely to report full-time employment compared with other fathers and were more likely to work part time.2 A possible explanation is that the burden of autism on the family economy is mediated by different employment choices made by caregivers in response to inadequate community resources.
Research has shown that parents of children with emotional and behavioral disorders often lack appropriate community-based services and resources needed to support work and family obligations.3 Families may have to accommodate work and family life to the behaviors and needs of the disabled child.4 In this context, it is not surprising that families with children with serious disabilities balance family and work responsibilities differently from other families.5,6
For preschool-aged children, a critical community resource is the availability and stability of child care services that meet both the child's needs and the parent's work schedule. Recent research has shown that that even publicly funded prekindergarten programs expel children with behavioral problems at higher rates than the public school system at older grades.7–9 The rates of expulsion for children with developmental problems from privately funded and operated child care centers are unknown but are believed to be higher, given that the private sector does not have the restrictions against expulsion that are present in the public sector.
Because of the national interest in understanding how child care problems interfere with work among low-income populations, the National Survey of Children's Health (NSCH) 2003 asked parents of young children whether child care problems had greatly affected their employment decisions in the last 12 months. Using this nationally representative data source, our objectives in this study were to describe the child care arrangements of children with autism compared with typically developing children and children at high risk for developmental problems and to determine whether families of preschool-aged children with autism were more likely to report that child care arrangements affected employment compared with families with typically developing children and children at high risk for developmental problems.
The NSCH is part of the State and Local Integrated Telephone Survey program conducted by the Centers for Disease Control and Prevention National Center for Health Statistics to track and monitor the health and well-being of children and adults at both the state and national levels. The survey was conducted between January 2003 and July 2004 (87% of the interviews were completed in 2003). The survey interviewed the adult in the household most knowledgeable about the sampled child's health, typically the mother. The response rate was 68.8%. 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. The NSCH completed 102353 interviews, averaging ∼2000 per state.
The questions regarding child care (section 6: early childhood) were only asked of parents of children who were ≤5 years of age. There were very few 0- to 2-year-old children diagnosed with autism. Thus, we restricted the sample to 3-, 4-, and 5-year-old children.
Autism was measured with the following yes-or-no question: “Has a doctor or health professional ever told you that [child] has any of the following conditions? Autism?” This question has been used as a measure for autism diagnosis in other national studies of children with autism by the Centers for Disease Control and Prevention and other researchers.10,11
Main Dependent Variable
Our measure of child care problems affecting employment was an affirmative response to the following yes-or-no question: “During the past 12 months did anyone in the family quit a job, not take a job, or greatly change your job because of problems with child care for [child]?”
Developmental delays were assessed with the Parent's Evaluation of Developmental Status (PEDS) questionnaire. The PEDS is designed to identify children who have or are at risk for developmental problems, including emotional and behavioral problems.12 Recent evaluation of the PEDS for the preschool population (2–4 years of age) showed the PEDS to have good psychometric properties.13 The PEDS classification of children at high risk (path A) for preschool-aged children identifies children whose parents had ≥2 concerns in 4 areas: global, expressive language, receptive language, and gross motor skills. For 5-year-olds, 2 additional areas of concern were added: fine motor skills and learning and/or school problems. Recent research has shown that the PEDS has good sensitivity but low specificity to detect ASDs in young children.14,15
Child Care Variables
Type of child care arrangement was measured by a series of 6 yes-or-no questions that asked, “if, during the last month, the child regularly attended” a child care center, family-based child care outside of the parent's home, in-house care by nonparents, Head Start, and/or nursery school, preschool, or kindergarten. We also created a “multiple child care arrangement” indicator that described whether, during the last month, the child regularly attended 0, 1, 2, or ≥3 different child care settings from the list above. Children who did not attend any of these settings regularly were considered to be in exclusive parental care. Instability of child care was measured by a response indicating ≥2 last-minute child care arrangements based on the following question: “During the past month how many times have you had to make different arrangements for childcare at the last minute because your usual plans changed due to circumstances beyond your control?”
Family structure was measured with a dichotomous 2-parent family indicator (2-parent households with biological, adoptive, or step parents versus single-parent families and all other family structures). Family poverty was coded by using the derived NSCH variable that used income and family size information to classify the household income above or below 200% of the 2003 poverty line on the basis of US Department of Health and Human Services guidelines. Parental education was coded using a derived variable that tracked the highest level of education in the household into 2 groups: high school graduate or less and more than high school education. Age, gender, and minority ethnicity (all groups except white, non-Hispanic) were measured using direct questions from the survey. All of the participants consented to the study. Additional information on variables and consent procedures related to this survey is available.16
Stata (Stata Corp, College Station, TX) was used to adjust for the complex sample design using Taylor approximations that provide the correct SEs, following NSCH guidelines.16
We used χ2 analyses for univariate comparisons and logistic regression for multivariate analyses. To graphically plot the multivariate logistic model, we created an index of covariates that summarized all of the risk factor information and plotted the logistic function over the range of the index of the covariates for each of the 3 groups. This procedure was based on the Stata vibl17 (visualizing binary logit models) package.
Prevalence of ASD
There were 82 children with autism in the sample for a population prevalence of 39 in 10000 (95% confidence interval [CI]: 26–51 in 10000). This is lower than the commonly accepted prevalence of ASD at 60 in 10000,11 but a lower prevalence was expected given that many children with ASD in the United States do not receive a diagnosis until after age 5 years.
The prevalence of ASD differed substantially by child care arrangement. Figure 1 displays the prevalence estimates of ASD by regularly attended child care arrangement. There were almost no children with diagnosed autism in exclusive parental care (5 in 10000 [95% CI: 0–11 in 10000]), whereas the prevalence among children who attend preschool, nursery school, and kindergarten was close to the expected 60 in 10000 (61 [95% CI: 40–81 in 10000]). Among children who attend Head Start, the prevalence was 89 in 10000 (95% CI: 40–139 in 10000).
Prevalence of High-Risk PEDS
There were 2027 children who screened positive for path A (high-risk) in the PEDS, for a population prevalence of 13.7% (95% CI: 12.8%–14.7%). These 2027 children represent 1720000 high-risk 3- to 5-year-old children nationally. As expected, 96% of children with an ASD diagnosis screened positive in the PEDS.
The number of children at high risk who were not diagnosed with ASD was 1955. These children formed the high-risk group for analyses. A few children with ASD had missing PEDS data and vice versa.
Classification of the Undiagnosed ASD Cases
Given the commonly accepted prevalence of ASD of 60 in 10000, we would have expected to observe approximately 128 case subjects in our data if all of the children had been diagnosed. We observed 82; thus, we likely were missing 44 children, or roughly one third of the children with ASD. On the basis of the sensitivity of the PEDS for autism,12,15 we would have expected 98% of those 44 children to screen positive in the PEDS, yielding an estimated 43 children in the high-risk group who had undiagnosed ASD. This constituted only 2% (43 of 1955) of the high-risk group. Thus, the high-risk group seems to represent a developmentally at-risk group for reasons other than diagnosable ASD. We expected only 1 undiagnosed children with ASD to be present in the typically developing group representing <0.01% (1 of 14245) of the group. Thus, the presence of undiagnosed children with ASD posed no threat to comparisons between the typically developing group and the ASD group. The 44 undiagnosed children with ASD are estimated to represent 26100 children aged 3 to 5 years old with undiagnosed ASD nationally.
Demographic and Household Characteristics of the Sample
Table 1 shows the demographic and household characteristics of the sample. Compared to the typically developing group, children with autism were more likely to be boys (81% vs 50%; P < .01). There were no statistically significant differences by age, race, parental education, poverty, family structure, or having no full-time worker in the household. Compared with the high-risk group, children with autism were more likely to be boys (81% vs 64%; P < .05) and white non-Hispanic (77% vs 54%; P < .01).
Finally, compared with the typically developing group, children at high risk were more likely to be 5-year-olds (42% vs 30%; P < .01), boys (63% vs 50%; P < .01), live in a household <200% of the federal poverty line (54% vs 40%; P < .01), and have no person in the household working full-time (14% vs 8%; P < .01). High-risk children were less likely to be white, non-Hispanic (54% vs 67%; P < .01), have a parent with more than a high-school education (58% vs 68%; P < .01), or live in a 2-parent household (67% vs 76%; P < .01).
Child Care Arrangements
Table 2 shows the child care arrangements of the 3 groups of children. Compared with typically developing children, children with ASD were less likely to be in exclusive parental care (3% vs 23%; P < .01) and more likely to attend preschool, nursery school, or kindergarten (91% vs 58%; P < .01) and Head Start (39% vs 14%; P < .01).
Compared with children at high risk, children with ASD were less likely to be in exclusive parental care (3% vs 19%; P < .01) and more likely to attend preschool, nursery school, or kindergarten (91% vs 60%; P < .01).
Compared with typically developing children, children at high risk were more likely to be in family-based care (32% vs 28%; P < .05) or Head Start (25% vs 14%; P < .01) and less likely to be in exclusive parent care (19% vs 23%; P < .05). They were also more likely to be in ≥3 child care settings in the previous month (28% vs 23%; P < .01) and to report ≥2 last-minute child care arrangements in the previous month (unstable care; 25% vs 18%; P < .01).
Child Care Problems and Employment: Univariate Analysis
Thirty-nine percent of parents of a child with ASD reported that child care problems for the child with ASD had greatly affected their employment decisions (quit a job, not taken a job, or greatly changed their job) in the previous year compared with 16% for parents of children at high risk and 9% for parents of typically developing children (Table 2). All 3 of the groups were significantly different from each other (P < .01).
Child Care Problems and Employment: Multivariate Logistic Model
Table 3 shows the odds ratios (ORs) and associated 95% CIs for the multivariate logistic model. Covariates included were the child's age, gender, and ethnicity; living in a 2-parent household; living in a household with an annual income <200% of the federal poverty line; and parental education.
Households in poverty were 2 times more likely to report that child care problems had had a major impact on employment decisions than comparable nonpoor households (OR: 2.02 [95% CI: 1.57–2.59]). If the child had minority ethnicity, the parents were also more likely to report that child care problems had greatly affected employment (OR: 1.28 [95% CI: 1.01–1.63]). Screening positive in the PEDS (high-risk group) was as powerful as living in poverty in terms of increasing the odds that child care problems affected employment (OR: 2.02 [95% CI: 1.54–2.63]). Having a child with ASD increased the odds of reporting that child care interfered with employment sevenfold (OR: 7.31 [95% CI: 3.44–15.54]), >3 times the effect of poverty. The high-risk and ASD groups were significantly different in multivariate analyses (design-corrected Wald test F = 10.34; P < .01).
Figure 2 shows the predicted probability that child care problems had affected employment decisions for the 3 groups of children on the basis of the multivariate logistic model. The index of covariates was formed by multiplying the logistic covariate coefficients by the value of the covariate variables. In this case, higher values of the covariate index indicated additional risk factors for child care problems (eg, poverty and minority ethnicity).
The predicted probability of child care problems affecting employment decisions rose nonlinearly as the covariate index increased for all 3 of the developmental groups but rose far more rapidly for children with autism and children at high risk than for typically developing children. The model predicted that, in the presence of other risk factors, such as poverty and minority ethnicity, families would have a more than proportional increase in the probability that child care problems would impact their employment decisions greatly. This was true for both the high-risk and ASD groups. Because the ASD group has ∼20% higher probability than the high-risk group even under the most favorable circumstances (no other risk factors), in the presence of socioeconomic risk factors, the model predicted that the majority of parents of children with ASD would experience child care problems that interfere with work.
To our knowledge, this is the first population-based study evaluating the impact of child care problems on the employment decisions of parents with preschool-aged children with ASDs in the United States. We found that families with preschool-aged children diagnosed with ASD use child care more than other families, and families with children diagnosed with ASD are more likely than other families to report that child care problems had greatly affected their employment decisions.
In this study, 97% of diagnosed preschool-aged children with ASD were cared for in community settings, often with multiple arrangements that typically included preschool, nursery school, or kindergarten (91%) and/or Head Start (39%). In contrast, very few children (3%) with ASD were cared for exclusively by their parents in the home. There were no significant differences in other types of child care (eg, family-based or child care center). These arrangements were accompanied by disruptions in employment for many families. We found that 39% of parents of children with diagnosed ASD reported that, in the previous year, someone in the family had quit a job, not taken a job, or greatly changed a job because of child care problems related to the child with ASD. These findings were substantiated in multivariate analyses controlling for several potentially confounding variables. In fact, the size of the effect was 3 times larger than the effect of poverty, suggesting that having a child with ASD may be the most important risk factor for child care problems affecting employment among these families.
These results should be interpreted in the context of the Individuals With Disabilities Education Act. The Individuals With Disabilities Education Act provides funds for services for disabled children in 2 broad categories: part B (ages 3–21 years) and part C (ages 0–3 years). Part B services are provided in the public school setting through an Individualized Education Plan, whereas part C services are often provided in the home through an Individualized Family Service Plan via community or county agencies.18 In practice, most preschool-aged children who do not attend preschool through the public school district do not have Individualized Education Plans and do not receive services. In this context, the findings that 91% of children with ASD attend preschool, nursery school, or kindergarten and 39% also attend Head Start likely represent the need for enrollment into school-based programs that provide free services for the affected child (eg, speech and occupational therapy). It is a parental response to federal disability policy that funnels assistance through the school venue.
Given the child care arrangements, it seems that parents of children with ASD are modifying employment to enable the typical child with ASD to attend school-related services and receive adequate wrap-around care. Doing so, however, must require substantial accommodation, because families with a child with ASD were 7 times more likely to report that child care problems affected employment than comparable families with typically developing children. These accommodations probably result in lower household income.1 Given that families of children with ASD face substantial additional health care expenses,19,20 the impact of child care problems on the economy of families with a child with ASD may be even more significant than for other families. We found that, in the presence of additional risk factors, the impact of child care problems on employment rises more than proportionally, representing a compounded burden for families already in need. This result is corroborated by previous research21,22 on child care and employment for families with children with disabilities. Although not the focus of this study, parents of other children at high risk seem to face similar problems to a lesser degree.
In this study, we found no statistically significant differences between children with ASD and other children in the instability of child care arrangement or multiple child care arrangements. However, in both cases the ASD group had comparable proportions to those of the high-risk group (which were significantly different from those of the typically developing group); thus, we likely lacked sufficient statistical power to show differences on these measures given the relatively small sample size of the ASD group.
The high prevalence of ASD in Head Start merits some comment. Because it was significantly higher than the commonly accepted prevalence of 1 in 150, it cannot be the result of enhanced developmental screening procedures alone. Head Start provides developmental screening, comprehensive services,23 does not expel children with behavioral problems, and can accept ≤10% of its students from higher income groups, particularly if they are disabled. Children with ASD represent a small proportion of the population (<1%), thus, they may find Head Start services available to them regardless of income. Because expulsions of preschool-aged children with behavioral problems in other child care settings are not uncommon,7,8 children with significant developmental problems may find in Head Start a particularly stable child care arrangement.
Finally, the prevalence of diagnosed autism in the United States for children aged 3 to 5 was 39 in 10000 cases, suggesting that one third of the group (26100 children nationally) had not yet been diagnosed. The American Academy of Pediatrics Council on Children With Disabilities recently released a policy statement recommending systematic developmental screening in primary care practices for early identification of developmental concerns, including ASD.24–27 Such a policy seems particularly pertinent in light of these data. Although some children may be identified over time through screening at child care settings, the setting with the fewest case subjects with autism was exclusive parental care. Thus, developmental screening at pediatric offices may be one of the few available venues to identify children with ASD who are not in community settings.
There are a few potential limitations in this study. Although parent report of ASD is viewed as a fairly reliable,11,28 we did not have access to medical charts or diagnostic reports, and we were limited by the way the ASD and child care questions were asked in the survey. In addition, we had a relatively small sample of children with ASD, and, thus, could not investigate further differences by subgroup. Lastly, we can only report on the questions asked in the survey and do not have detailed information regarding specific barriers to child care services.
Developmental problems in general and ASD in particular are associated with higher use of child care services and higher probability that child care problems affect employment among families with preschool-aged children in the United States. In addition, systematic developmental screening for ASD is needed for young children in child care settings and in pediatric practices.
Drs Montes and Halterman were funded in part by US Administration of Children and Families grant 90YE0102/01 and the Rochester Area Community Foundation.
We thank Connie Valk (early care and education specialist, Children's Institute) and Paula Boughton (education disabilities coordinator, ABC HeadStart) for consultation regarding child care settings.
- Accepted January 17, 2008.
- Address correspondence to Guillermo Montes, PhD, Children's Institute, 271 N Goodman St, Suite D103, Rochester, NY 14607. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject
Childhood autism is associated with increased health care, educational, and behavioral expenses and with earning 14% less income.
What This Study Adds
Childhood autism is associated with increased enrollment into school-based settings for preschool-aged children. Parents are 7 times more likely to report that child care problems substantially affected employment decisions if they have a child with ASD. Approximately one third of preschool-aged (3 to 5 years) children with ASD are undiagnosed.
- ↵Montes G, Halterman JS. Association of childhood autism spectrum disorders and loss of family income. Pediatrics.2008;121 (4). Available at: www.pediatrics.org/cgi/content/full/121/4/e821
- ↵Rosenzweig JM, Brennan EM, Ogilvie AM. Work-family fit: voices of parents of children with emotional and behavioral disorders. Social Work.2002;47 (4):415– 424.
- ↵Gilliam WS. Prekindergarteners Left Behind: Expulsion Rates in State Prekindergarten Systems. New Haven, CT: Yale University Child Study Center;2005
- ↵Grannan M, Carlier C, Cole CE. Early Childhood Care and Education Expulsion Prevention Project. Southgate, MI: Downriver Guidance Clinic Department of Early Childhood Programs;1999
- ↵Glascoe FP. Parents' evaluation of developmental status: how well do parents' concerns identify children with behavioral and emotional problems? Clin Pediatr (Phila).2003;42 (2):133– 138
- ↵Young L, Pinto-Martin JA, Warszawa A, Giarelli E, Levy SE. Comparison of a general developmental screening tool and an autism specific screening tool in autistic spectrum disorder (ASD) assessment. J Dev Behav Pediatr.2006;27 (5):431
- ↵Glascoe FP, Macias M, Wegner LM. Can parents' concerns identify children likely to have autism-spectrum disorder? Presented at the Pediatric Academic Societies Annual Meeting; May 3–6,2003; Seattle, WA
- ↵Blumberg SJ, Olson L, Frankel MR, Osborn L, Srinath KP, Giambo P. Design and operation of the National Survey of Children's Health, 2003. Vital Health Stat.2005;1 (43):1– 24
- ↵Mitchell M, Chen X. vibl: visualizing binary logistic models. UCLA Academic Technology Services. Available at: www.ats.ucla.edu/stat/stata/seminars/stata_vibl/viblm.htm. Accessed October 10,2007
- ↵Oser C, Cohen J. Improving part C early intervention: using what we know about infants and toddlers with disabilities to reauthorize part C of IDEA. Washington DC: Zero to Three Policy Center,2003
- ↵Ward H, Atkins J, Herrick A, Morris P. Child Care and Children With Special Needs: Challenges for Low Income Families. Portland, ME: Edmund S. Muskie School of Public Service, University of Southern Maine;2004
- ↵Ward H, Morris L, Atkins J, Herrick A, Morris P, Oldham E. Child Care and Children With Special Needs: Challenges for Low Income Families. Portland, ME: Edmund S.Muskie School of Public Service, University of Southern Maine;2006
- ↵Irish K, Schumacher R, Lombardi J, for the Center for Law and Social Policy. Head Start Comprehensive Services: A Key Support for Early Learning for Poor Children. Washington, DC: Center for Law and Social Policy;2004
- ↵Gupta VB, Hyman SL, Johnson CP, et al. Identifying children with autism early [published correction appears in Pediatrics. 2007;119(4):867]? Pediatrics.2007;119 (1):152– 153
- Williams J, Brayne C. Screening for autism spectrum disorders: what is the evidence? Autism.2006;10 (1):11– 35
- ↵Montes G, Halterman JS. Psychological functioning and coping among mothers of children with autism: a population-based study. Pediatrics.2006;119 (5). Available at: www.pediatrics.org/cgi/content/full/119/5/e1040
- Copyright © 2008 by the American Academy of Pediatrics