585563

research-article2015

ECXXXX10.1177/0014402915585563Exceptional ChildrenMarshall et al.

Article

Parent Concern and Enrollment in Intervention Services for Young Children With Developmental Delays: 2007 National Survey of Children’s Health

Exceptional Children 2016, Vol. 82(2) 251­–268 © 2015 The Author(s) DOI: 10.1177/0014402915585563 ec.sagepub.com

Jennifer Marshall1, Russell S. Kirby1, and Peter A. Gorski2

Abstract This study sought to address underenrollment and late entry to early intervention by identifying factors associated with parental concern and services for developmental delays. The authors analyzed responses from 27,566 parents of children from birth to age 5 from the 2007 National Survey of Children’s Health to quantify and to identify factors associated with developmental concerns and enrollment in public intervention or therapy. Developmental concerns were common among parents from all backgrounds, increasing as children approach preschool age and particularly among children with poor health and those with non-English home language. However, enrollment in intervention is low. Nearly 40% of parents reported one or more concerns, yet 5% of children were enrolled in public intervention or therapy. Multirace or Black race, non-English home language, low income, and private or no insurance were associated with lower odds of services enrollment. Primary health care provider and parent involvement were associated with higher likelihood of parent-reported concern and services. National surveys provide the means to collect information directly from a nationally representative sample of parents on important topics related to their children’s health and development. We examined parents’ reports of their developmental concerns and enrollment in services for children from birth to age 5 in order to identify sociodemographic factors that may be associated with early recognition and intervention for developmental delays. Identifying young children with developmental delays in the United States is particularly challenging because this population spans all stages of recognition, help-seeking, and service enrollment. Current estimates of developmental delays in children under age 5 range from 15% to 17% (Boyle et al., 2011; Rosenberg, Zhang, &

Robinson, 2008), and the percentage of parents with concerns and this age group ranges from 9% to 23% depending on how the question is framed (Glascoe, 1997; Smith, Akai, Klerman, & Keltner, 2010). As a result, the prevalence of children with delays is usually estimated from surveys, clinical samples, or intervention programs, which may exclude

1

University of South Florida The Children’s Trust

2

Corresponding Author: Jennifer Marshall, PhD, MPH, Department of Community and Family Health, College of Public Health, University of South Florida, 13201 Bruce B Downs Blvd., MDC56, Tampa, FL 33612, USA. E-mail: [email protected]

252 children who have concerned parents but are not enrolled in intervention services (Roux et al., 2012; Smith et al., 2010). Although developmental disabilities can be detected in children under age 2 (Valicenti-McDermott, Hottinger, Seijo, & Shulman, 2012), many delays go unrecognized by professionals, such as child care providers or pediatricians (Chung et al., 2011; Glascoe, 2013; Smith et al., 2010). However, some delays in young children may have elicited parent concerns, the majority of which have a sound basis (Glascoe, 1997).

Multirace or Black race, non-English home language, low income, and private or no insurance were associated with lower odds of services enrollment. Primary health care provider and parent involvement were associated with higher likelihood of parent-reported concern and services. Early intervention services for children with developmental delays can support development during important windows of opportunity, correct maladaptive patterns of development, and improve school readiness (Pinto-Martin, Dunkle, Earls, Fliedner, & Landes, 2005; Ramey & Ramey, 2004). Unfortunately, high rates of underenrollment and 1- to 3-year delays in entry to the U.S. public early intervention system are commonplace for a number of reasons (Pinto-Martin et al., 2005; Rosenberg, Robinson, Shaw, & Ellison, 2013). Differences among state intervention program eligibility criteria, funding for child-find identification activities, and other early childhood programs have an impact on developmental screening, referral, and enrollment. Health care providers may not refer young children who show signs of developmental delays or may have limited information on where to refer (Guevara et al., 2013; Hix-Small, Marks, Squires, & Nickel, 2007). Prior research has shown that parents who do have concerns may search for assessment or

Exceptional Children 82(2) intervention services for months to years (Seligman & Darling, 2007), and some parents who are aware of developmental delays in their children may not seek services due to factors such as the belief that the child will grow out of it (Shevell, 2008) or perceived practical or cultural barriers to accessing formal services (Guevara et al., 2013; Hix- Small et al., 2007; Pinto-Martin et al., 2005). Thus, the majority of children under age 5 with developmental delays do not access developmental services, and parents who do pursue services may enroll their children in public intervention programs, private therapy services, or a combination of the two. The socioecological model has been increasingly used in public health, as it has been in early intervention, because of its utility in identifying the multilevel influences on child and adult health to identify leverage points for improvement in recognition and access to services (Bronfenbrenner, 1979). Further, research using the life-course perspective has demonstrated that sociodemographic disparities in early childhood, exacerbated by conditions at all levels of the socioecological model, have a lasting and cumulative impact on adult health, wealth, and well-being (Children’s Defense Fund, 2014; Lu & Halfon, 2003). In keeping with a socioecological approach, this crosssectional exploratory study utilized a large, nationally representative data set (the 2007 National Survey of Children’s Health [NSCH]; Blumberg et al., 2012) to examine several factors within the child, parent, family, and community domains that have been found in the literature to be related to early intervention access and utilization. The outcome variables we examined included recognition (operationalized as parent concern) and services enrollment (operationalized as receiving services from public intervention programs funded through the Individuals With Disabilities Education Act or from private therapy providers) for developmental delays in children from birth to age 5. By conducting multivariable analyses of data on a large and comprehensive set of key variables, as reported by the parents of young children across the country, we sought to

253

Marshall et al. address the following research questions: (a) What individual, interpersonal, and community-level factors contribute to parent recognition of developmental delays in their young children? and (b) What are the individual, interpersonal, and community-level factors that most contribute to parents’ help-seeking for developmental concerns?

Method Following determination of exempt status by the University of South Florida Internal Review Board, this study examined data from a nationally representative sample of racially, ethnically, and geographically diverse parents of 27,566 children from birth to age 5 participating in the 2007 NSCH, a representative sample survey that collects parent-reported information on the physical and mental health status, health care quality and access, and the family, neighborhood and social context (Blumberg et al., 2012). NSCH questions related to developmental concerns are adapted from the Survey Parents’ Evaluation of Developmental Status (PEDS; PEDSTest.com), which, in contrast with the clinical version of the PEDS (Glascoe, 2013), asks 12 closedended questions regarding whether the parent has an overall concern about the child’s learning, development, or behavior and also specifically about concerns regarding the child’s fine motor, gross motor, receptive language, expressive language, self-help, behavior, social-emotional, and, for older children, preschool and school skills. The survey also collects data on many aspects of child and family health and functioning, access and utilization of medical and community services, and family structure and demographics. Several variables are available individually and also in composite scales for parent concern, developmental problems, at-risk status, health insurance coverage, primary pediatric care or medical home, child care, developmental screening, and social support. Comparison groups in this study include NSCH parents who have developmental concerns versus those without and parents with children

enrolled in public intervention (“child has an individualized family support plan [IFSP] or individualized education plan [IEP]”) or private therapy (NSCH variable asking parent if child “qualified for use of special therapies”) versus those without. Independent variables were organized into child, parent, family, and community domains to examine their associations with the concern and intervention outcomes.

Analysis We examined all variables related to child and family characteristics (child sex, age, health status, language spoken at home, maternal age, education, income, marital status), parental involvement (read to child or take on outing), access to services (medical home, insurance coverage, utilization of specialty care, number of doctor visits, developmental screening, sources of medical care), child care utilization and type, and family structure (number of adults and siblings in the household) for construct validity. Quantitative analysis was conducted using SAS 9.3 survey statistics procedures, which account for the complex sampling utilized in the NSCH (Blumberg et al., 2012). We examined data for missing values and outliers; variables with missing data above 5% (maternal education, 5.19%; age, 5.04%; self-help concern, 14.44%; and academic concern, 28.58%) were formatted to include a missing stratum in the analysis. A composite concern variable was constructed from the general and specific concern questions, a composite “racelanguage” variable combined race and ethnicity and home language to account for interactions, and maternal age and number of days read to child were grouped into categories to simplify interpretation of results. We examined the characteristics of subgroups of parents with and without developmental concerns using PROC SURVEYFREQ to generate frequencies and weighted distributions, and chi-square tests assessed statistically significant differences (two tailed using alpha < .05) between groups on specific variables.

254

Exceptional Children 82(2)

Therapy 2%

IFSP/IEP 4%

Concerns 39%

No IFSP/IEP or therapy 33%

No concerns 61%

Figure 1.  Services enrollment among parents reporting concern for children from birth to age 5. IFSP = individualized family services plan; IEP = individualized education program.

Logistic regression models were estimated using maximum likelihood estimation, incorporating the sample design into the analysis. Adjusted odds ratios (aOR) and corresponding 95% confidence intervals were calculated; statistical tests were two tailed, using alpha < .05 to determine significance. Three separate series of univariate regression analyses examined the outcomes concern, enrollment in public services (IFSP/IEP), and enrollment in private services (therapy). Regression models were also tested for interactions among covariates; variables that interacted with more predictive variables were removed from the model, and more specific variables were used (i.e., children-with-special-health-care-needs [CSHCN] variable replaced with health status, as the CSHCN definition includes diagnosed developmental disability). The race and ethnicity and language variables exhibited multicollinearity, resulting in misspecified models when we included each in the multivariable models. We had several choices: (a) remove one of these variables from the model, (b) combine the two variables into a single variable by cross-classifying, or (c) make the models even more complex by including interaction terms. We did not want to lose the information about language spoken in the household and felt that the strategy of combining the two variables into a single crossclassified variable was the best solution.

Univariate logistic regression identified several characteristics in child, parent, family, and community domains with statistically significant associations with increased or decreased odds of concern and services enrollment. Multivariable logistic regression models with all covariates included were used to estimate aORs and 95% confidence intervals for parent concern, enrollment in public intervention or special education (IFSP/IEP), and enrollment in private therapy.

Rates of concern and enrollment in public early intervention varied by state; the percentages of parents reporting developmental concern ranged from 27% to 46% (M = 35.88%), and enrollment in public intervention (affected by states’ varied eligibility criteria) ranged from less than 1% to 9% (M = 1.96%).

Results Figure 1 and Table 1 present the distribution of respondents by developmental concern and enrollment in public intervention and therapy services. Rates of concern and enrollment in public early intervention varied by state; the percentages of parents reporting developmental concern ranged from 27% to 46% (M = 35.88%), and enrollment in public intervention (affected

255

Child domain   Child’s age   40   Missing

Variable

6,544 657 * 234 3,574 4,478 1,100 671

** 65 113 158 220 288 307 ** 751 399 ** 472 337 225 95 22

n

2.98 (0.28) 17 38.82 (0.95) 343 45.48 (0.94) 506 8.24 (0.50) 193 4.47 (0.34) 92

2.92 (0.48) 38.58 (1.26) 42.38 (1.21) 10.14 (0.80) 5.96 (0.65)

485 6,383 8,288 1,550 803

64.05 (0.95) 779 4.39 (0.34) 90

54.20 (1.28) 12,856 6.48 (0.68) 728

4,674 4,719 3,801 4,459 4,403 4,234

n

2,155 4,279

2.74 (1.83) 691 35.24 (3.16) 9,557 41.03 (3.12) 12,226 14.87 (2.30) 2,448 6.12 (1.16) 1,368

59.43 (3.23) 18,575 6.64 (1.37) 1,281

11.12 (2.44) 22.82 (2.76)

38.46 (3.19) 18,539 31.36 (3.24) 5,260 19.56 (2.23) 2,087 9.16 (1.49) 352 1.46 (0.47) 46

2.96 (0.25) 38.83 (0.78) 44.48 (0.77) 8.75 (0.44) 4.98 (0.35)

60.41 (0.80) 5.11 (0.35)

12.25 (0.60) 22.23 (0.70)

66.41 (0.77) 20.96 (0.63) 10.57 (0.58) 1.87 (0.30) 0.18 (0.04)

48.83 (0.76) 47.48 (0.76)

17.43 (0.60) 17.76 (0.59) 15.31 (0.56) 16.34 (0.47) 16.18 (0.57) 16.97 (0.60)

% (SE)

No IEP/IFSP N = 29,290 (96.12%)

65.29 (3.21) 13,416 34.71 (3.21) 12,857

6.22 (1.64) 11.58 (2.29) 12.02 (1.76) 15.69 (2.33) 24.05 (2.43) 30.45 (3.23)

% (SE)

IEP/IFSP N = 1,151 (3.88%)

10.57 (0.70) 73 20.98 (0.81) 209

72.78 (0.87) 18.28 (0.71) 8.08 (0.64) 0.78 (0.11) 0.08 (0.03)

54.15 (1.27) 13,270 26.40 (1.13) 3,037 15.62 (1.02) 1,035 4.36 (0.72) 150 0.48 (0.09) 13

1,208 2,717

29.90 (0.68) 32.48 (0.72)

8,567 8,928

21.53 (0.63) 16.76 (0.59)

22.63 (0.78) 18.42 (0.74) 15.39 (0.72) 14.87 (0.71) 14.38 (0.69) 14.32 (0.60)

% (SE)

3,973 3,241 2,457 2,675 2,587 2,576

n

No concern N = 17,509 (61.73%)

7.93 (0.82) 16.07 (0.90) 14.82 (0.81) 18.61 (0.92) 19.92 (0.93) 22.65 (1.16)

% (SE)

** 1,052 14.97 (1.04) 1,804 24.35 (1.17)

** 796 1,612 1,521 2,019 2,127 1,982 ** 5,676 4,376 ** 5,801 2,593 1,304 300 56

n

Concern N = 10,057 (38.27%)

522 794 * 8 245 342 142 57

62 155

** 56 98 112 163 186 179 ** 529 265 ** 201 241 219 103 30

n

4,713 4,755 3,866 4,531 4,528 4,379

n

.84 (0.39) 711 33.92 (3.10) 9,712 44.28 (3.60) 12,424 15.23 (2.19) 2,508 5.73 (1.33) 1,417

59.81 (3.35) 18,878 6.30 (1.50) 1,330

12.63 (2.36) 2,198 21.26 (2.62) 4,366

21.59 (2.97) 18,870 31.32 (3.47) 5,389 27.71 (2.94) 2,120 15.84 (2.35) 347 3.54 (0.89) 39

(continued)

3.02 (0.26) 38.86 (0.77) 44.30 (0.76) 8.80 (0.44) 5.03 (0.33)

60.29 (0.79) 5.16 (0.34)

  12.25 (0.60) 22.30 (0.69)

66.42 (0.76) 21.12 (0.63) 10.52 (0.57) 1.79 (0.29) 0.15 (0.04)

50.67 (0.75) 49.33 (0.76)

17.29 (0.59) 17.61 (0.59) 15.23 (0.55) 16.25 (0.57) 16.31 (0.57) 17.31 (0.60)

% (SE)

No therapy N = 26,772 (97.43%)

67.89 (3.16) 13,714 32.11 (3.16) 13,039

6.34 (1.49) 13.99 (2.32) 12.85 (0.16) 18.03 (3.26) 23.74 (2.69) 25.06 (3.07)

% (SE)

Therapy N = 794 (2.57%)

Table 1.  Frequency of Concern, IEP/IFSP, and Therapy Enrollment by Child, Parent, Family, and Community Characteristics.

256

 Child race or ethnicity   by home language   English    White    Black    Hispanic    Multiracial    Other   Non-English    White    Black    Hispanic    Multiracial    Other Family domain   Age position (siblings)   Only child   Oldest child   Second oldest   Third oldest   Fourth oldest   Household income   ≤100% FPL   100%–199%   200%–399%   ≥400% FPL  Parent involvement   (days read to child   per week)

Variable

Table 1. (continued)

N/A

n

* 3,985 1,456 3,041 985 368 ** 1,814 1,995 3,193 3,055 ** 7,204 2,148 5,421 1,980 756 2,119 2,953 5,733 6,704

24.11 (1.10) 24.57 (1.19) 27.77 (1.05) 23.55 (1.01)

132 32 996 15 196

24.93 (0.91) 16.10 (0.88) 34.36 (1.19) 18.12 (1.18) 6.50 (0.71)

1.09 (0.32) 0.26 (0.07) 17.87 (1.24) 0.18 (0.10) 3.55 (0.68)

19.12 (0.87) 20.47 (0.77) 29.83 (0.82) 30.58 (0.85)

24.49 (0.68) 12.73 (0.58) 33.99 (0.89) 20.95 (0.92) 7.8 (0.62)

408 191 364 142 46 * 195 252 385 319 *

8 0 35 0 9

0.73 (0.10) 0.26 (0.15) 10.40 (0.80) 0.04 (0.01) 1.50 (0.29)

% (SE)

72 27 999 14 194

n

n

195 59 1,942 29 376

19.56 (2.24) 28.68 (3.00) 29.77 (2.84) 22.00 (3.02)

3,697 4,664 8,507 9,422

20.77 (2.34) 10,815 19.94 (2.84) 3,439 34.32 (3.05) 8,125 18.27 (2.63) 2,833 6.70 (1.54) 1,078

0.59 (0.42) — 6.60 (2.16) — 0.77 (0.34)

21.05 (0.73) 21.72 (0.68) 29.04 (0.66) 28.19 (0.67)

24.78 (0.57) 13.81 (0.50) 34.12 (0.73) 19.94 (0.75) 7.35 (0.48)

0.88 (0.14) 0.27 (0.10) 13.51 (0.71) 0.10 (0.04) 2.34 (0.33)

53.09 (0.78) 12.20 (0.46) 9.38 (0.54) 5.19 (0.40) 2.97 (0.25)

% (SE)

No IEP/IFSP N = 29,290 (96.12%)

58.30 (3.29) 16,587 13.38 (2.05) 2,248 10.45 (2.39) 2,043 6.59 (1.63) 1,434 3.02 (0.62) 926

% (SE)

IEP/IFSP N = 1,151 (3.88%)

57.61 (0.98) 746 11.27 (0.55) 98 9.61 (0.67) 101 5.55 (0.57) 80 2.93 (0.04) 54

% (SE)

No concern N = 17,509 (61.73%)

5,693 46.19 (1.22) 11,686 1,057 13.86 (0.78) 1,305 827 9.14 (0.87) 1,332 580 4.76 (0.40) 941 418 3.03 (0.36) 573

N/A

n

Concern N = 10,057 (38.27%)

141 164 267 222 *

320 129 241 75 29

7 0 37 1 6

501 61 76 52 40

N/A

n

n

20.83 (2.56) 21.86 (2.71) 32.86 (0.313) 24.45 (3.58)

3,792 4,784 8,659 9,537

25.60 (2.62) 10,961 17.99 (2.87) 3,511 32.88 (3.03) 8,283 17.01 (3.46) 2,916 6.52 (1.83) 1,101

0.88 (0.63) 197 — 59 8.35 (1.98) 1,958 0.10 (0.10) 28 0.64 (0.31) 384

(continued)

  24.63 (0.56) 13.91 (0.50) 34.16 (0.73) 19.94 (0.74) 7.35 (0.48)   21.04 (0.69) 22.04 (0.68) 28.94 (0.66) 27.98 (0.66)

  53.12 (0.78) 12.21 (0.46) 9.38 (0.54) 5.30 (0.40) 2.96 (0.25)   0.87 (0.14) 0.27 (0.10) 13.37 (0.70) 0.09 (0.04) 2.33 (0.32)

% (SE)

No therapy N = 26,772 (97.43%)

57.82 (3.57) 16,878 14.02 (2.71) 2,301 11.52 (2.83) 2,083 3.24 (0.70) 1,469 3.45 (0.74) 951

% (SE)

Therapy N = 794 (2.57%)

257

3,999 12,292 1,058 **

41.05 (1.29) 50.98 (1.27) 7.97 (0.68)

90.80 (0.90) 9.20 (0.90)

42.32 (1.25) 30.35 (1.09) 15.50 (1.01) 11.83 (0.78)

9,400 634

3,884 3,556 1,372 1,244

16,837 653 * 7,534 5,947 2,421 1,603

8,578 8,191

45.02 (1.24) 50.91 (1.25)

4,940 4,753 ** 212 5,999 711

822 315 ** 478 607 50 48.38 (3.27) 47.25 (3.26) 4.36 (0.98)

67.40 (3.16) 31.82 (3.19)

48.03 (0.96) 28.33 (0.80) 14.19 (0.66) 9.46 (0.61)

388 36.69 (3.03) 503 37.59 (3.17) 113 9.95 (1.91) 147 15.78 (2.59)

95.32 (0.43) 1102 92.79 (2.15) 4.68 (0.43) 48 7.21 (2.15)

30.82 (0.94) 61.33 (0.98) 7.86 (0.61)

49.65 (0.94) 49.27 (0.95)

**

25,028 1,227 ** 10,970 8,972 3,657 2,686

6,668 17,646 1,712

12,631 12,582

67.67 (0.30) 19,971 12.04 (2.51) 2,338 16.52 (2.08) 3,322 3.78 (0.90) 484

74.69 (0.86) 805 10.75 (0.71) 95 12.89 (0.59) 192 1.67 (0.23) 51

n

2.15 (0.55) 1,465 23.55 (3.43) 4,909 74.30 (3.41) 19,784

% (SE)

10.89 (1.62) 1,629 68.12 (3.19) 20,224 21.00 (3.11) 4,352

n

6.72 (0.42) 122 75.25 (0.88) 821 18.04 (0.84) 202

% (SE)

85 553 151

32 116 641

n

46.20 (0.78) 28.77 (0.66) 14.88 (0.58) 10.15 (0.49)

93.62 (0.45) 6.38 (0.45)

760 33 * 277 336 77 104

593 196 ** 57.86 (0.80) 362 34.09 (0.79) 398 8.02 (0.48) 20

45.12 (0.77) 50.69 (0.78)

**

n

7.63 (0.37) 73.37 (0.73) 19.02 (0.69)   72.24 (0.71) 11.71 (0.59) 14.27 (0.50) 1.78 (0.17)

37.84 (3.25) 11,141 46.05 (0.78) 39.36 (3.55) 9,167 28.83 (0.65) 8.56 (1.68) 3,716 14.85 (0.58) 14.24 (2.54) 2,743 10.27 (0.49)

95.58 (1.21) 25,477 93.54 (0.45) 4.42 (1.21) 1,254 6.46 (0.45)

46.93 (3.47) 6,849 34.41 (0.78) 50.14 (3.52) 17,893 57.56 (0.79) 2.93 (0.97) 1,749 8.03 (0.47)  

71.56 (3.42) 12,925 45.35 (0.77) 27.74 (3.42) 12,748 50.48 (0.77)

67.61 (3.23) 20,300 12.46 (2.34) 2,382 17.34 (2.55) 3,402 2.30 (0.86) 512

9.64 (1.88) 1,677 70.00 (3.07) 20,563 20.37 (2.67) 4,445

7.68 (0.47) 23.11 (0.69) 69.21 (0.76)  

% (SE)

No therapy N = 26,772 (97.43%)

3.21 (0.83) 1,484 16.49 (2.54) 5,001 80.30 (2.64) 20,148

% (SE)

Therapy N = 794 (2.57%)

72.33 (0.72) 547 11.72 (0.60) 66 14.23 (0.51) 142 1.72 (0.17) 28

7.53 (0.37) 73.56 (0.74) 18.91 (0.69)

7.77 (0.48) 22.90 (0.70) 69.33 (0.76)

% (SE)

No IEP/IFSP N = 29,290 (96.12%)

807 9.23 (0.66) 955 7,345 70.17 (1.21) 13,771 1,858 20.69 (1.14) 2,738 ** 81 67.98 (1.19) 13,766 1,039 13.31 (1.00) 1,409 1,592 16.70 (0.87) 1,952 273 2.01 (0.22) 267

n

IEP/IFSP N = 1,151 (3.88%)

8.01 (0.60) 33 19.72 (0.76) 179 72.27 (0.88) 935

% (SE)

No concern N = 17,509 (61.73%)

522 6.84 (0.70) 994 2,139 28.16 (1.25) 2,978 7,331 65.00 (1.29) 13,458 *

n

Concern N = 10,057 (38.27%)

Note. IEP/IFSP = individualized education program or individualized family support plan; FPL = federal poverty level. Rao-Scott chi-square: *p ≤ .05, **p ≤ .0001. N/A = statistics could not be calculated as one cell contains n = 0.

  0 days   1–3 days   4–7 days  Number of adults in   the household   1   2   ≥3   Marital status   Married   Cohabiting   Neither   No parents Community domain  Provider ask about  concerns   Yes   No   Insurance type   Public   Private   None  Usual source of   medical care   Yes   No   Child care   Neither type   Nonrelatives   Relatives   Both types

Variable

Table 1. (continued)

258 by states’ varied eligibility criteria) ranged from less than 1% to 9% (M = 1.96%). Bivariate analysis identified several statistically significant differences between concern and no-concern groups in child age, sex, health status, age position (relative to siblings), frequency read to by an adult, maternal education, age, child race or ethnicity, primary language, marital status, family income, and number of adults in the household (Table 1). Compared by receipt of public and private developmental services, group differences were statistically significant on child age, sex, health status, and number and type of concerns but not by age position relative to siblings (Table 1). These group differences were not statistically significant by maternal or family characteristics, except for differences between groups in maternal age (for therapy vs. no-therapy groups) and frequency read to by an adult (for both service groups). Parents of children with public or private services differed significantly from parents reporting no services in insurance type, child care type, and whether the provider asked about concerns.

Concern Results of multivariable logistic regression models, including confidence intervals, are found in Table 2 and summarized in Table 3. Among the variables significantly associated with whether the family had concerns about child development, many were child variables. Likelihood of developmental concern was increased by age of child (ranging from aOR = 2.62 at age 1 to aOR = 4.75 at age 5), male sex (aOR = 1.40), and poorer reported health compared to excellent health (aOR ranged from 1.81 for “very good” to 7.70 for “poor” health). Likelihood of concern was also greater among children of Hispanic or non-White parents in a non-English-speaking home (aOR = 1.59 for Hispanic and 3.63 for multiracial), parents who read to the child moderately (1–3 days per week compared to none, aOR = 1.49), and parents who utilized relative child care compared to no child care (aOR = 1.31). Children from families with Missing values for maternal education had increased odds

Exceptional Children 82(2) (aOR = 2.18) for developmental concern. Increased number of older siblings was associated with a 20% to 40% decreased likelihood of parent-reported developmental concerns (one older sibling, aOR = 0.80; two older siblings, aOR = 0.67; three older siblings, aOR = 0.60). Parents who reported having a usual source of medical care were less likely to report concerns (aOR = 0.56), but those whose health care provider asked about concerns had increased likelihood of reporting concerns (aOR = 1.19). Family income, marital status, maternal age, and education were controlled for in the multivariable model but did not show an association with increased odds of reporting developmental concern. In the univariate models, but not in the multivariable model, likelihood of concern was statistically significantly higher among English-speaking Black parents (aOR = 1.58), children with younger siblings (aOR = 1.24), and children in nonrelative care (aOR = 1.19) or both nonrelative and relative care (aOR = 1.38) and was statistically significantly lower among households with two adults (aOR = 0.69) and families with private insurance (aOR = 0.62).

Public Intervention Services (IFSP or IEP) The odds of having a child with an IFSP or IEP were greater among parents of male children (aOR = 1.67); with relatively poorer reported health (“very good,” aOR = 1.61; “good,” aOR = 2.10; compared to “excellent”); who read to the child more (1–3 days, aOR = 3.89; 4–7 days, aOR = 4.60); had “any” concern about learning, development, or behavior (aOR = 4.18) or specific concerns with speech (aOR = 4.63), receptive language (aOR = 1.97), selfhelp (aOR = 1.63), or fine motor (aOR = 2.34) skills; and had a health care provider who asked about concerns (aOR = 2.09). In addition, families at 100% to 199% of federal poverty level (FPL) were more likely to report an IFSP or IEP compared to those below poverty level (aOR = 1.91). Parents who had private (aOR = 0.44) or no insurance (aOR = 0.33) compared to Medicaid were less likely to have an IFSP or IEP. Further,

259

Marshall et al.

Table 2.  Adjusted Odds Ratios (aOR) for Parent Concern, IFSP/IEP, and Therapy by Selected Child, Parent, Family, and Community Characteristics, Multivariable Model.

  Variable Child domain   Child’s age   40

Model 1: Concern

Model 2: IFSP/IEP

Model 3: Therapy

N = 25,552

N = 24,116

N = 24,202

aOR

95% CI

aOR

95% CI

aOR

Ref 2.62* 2.87* 4.18* 4.13* 4.75*

Ref [2.03, 3.38] [2.22, 3.72] [3.20, 5.46] [3.14, 5.44] [3.67, 6.15]

Ref 0.54 0.79 0.78 1.32 1.49

Ref [0.12, 2.46] [0.17, 3.76] [0.16, 3.72] [0.29, 6.07] [0.31, 7.10]

Ref 0.60 1.02 1.33 1.54 1.40

1.40* Ref

[1.23, 1.59] Ref

1.67* Ref

[1.17, 2.38] Ref

2.44* Ref

Ref 1.81* 2.11* 5.88* 7.70*

Ref [1.56, 2.10] [1.62, 2.76] [3.86, 8.96] [3.17, 18.74]

Ref 1.61* 2.10* 1.82 2.96

Ref Ref [1.07, 2.41] 3.72* [1.26, 3.51] 8.53* [0.88, 3.75] 15.82* [0.62, 14.15] 53.31*

Ref 0.87 4.63* 0.80 1.97* 0.97 1.36 1.63* 0.32 2.34* 0.73 4.18*

Ref [0.60, 1.26] [3.00, 7.15] [0.51, 1.25] [1.12, 3.46] [0.61, 1.55] [0.66, 2.80] [1.03, 2.59] [0.07, 1.46] [1.44, 3.80] [0.43, 1.24] [2.88, 5.97]

Ref 0.95 3.38* 0.60 1.47 0.92 1.93 2.36* 0.31 1.44 3.10* 6.22*

Not Not included included in Model 1 in Model 1

1.03 Ref 0.93 2.18*

[0.78, 1.38] Ref [0.78, 1.12] [1.17, 4.08]

0.83 Ref 1.04 1.15

[0.53, 2.04] Ref [0.74, 1.80] [0.28, 2.50]

1.93* Ref 1.26 0.26

Ref 0.86 0.94 1.13

Ref [0.52, 1.41] [0.56, 1.55] [0.29, 1.10]

Ref 0.56 0.56 0.71

Ref [0.09, 3.73] [0.09, 3.70] [0.10, 5.14]

Ref 7.00* 9.77* 11.08*

95% CI   Ref [0.16, 2.62] [0.24, 4.81] [0.32, 6.51] [0.36, 7.04] [0.31, 5.86]   [1.66, 3.60] Ref   Ref [2.38, 5.82] [5.27, 13.83] [7.74, 32.32] [12.42, 228.91]   Ref [0.59, 1.53] [2.21, 5.17] [0.32, 1.11] [0.75, 2.88] [0.49, 1.72] [0.96, 3.89] [1.31, 4.25] [0.07, 1.41] [0.80, 2.61] [1.78, 5.41] [4.15, 9.34]   [1.01, 3.70] Ref [0.89, 2.35] [0.02, 2.84]   Ref [2.04, 24.05] [2.75, 34.76] [2.75, 44.72] (continued)

260

Exceptional Children 82(2)

Table 2. (continued)

 

Model 1: Concern

Model 2: IFSP/IEP

Model 3: Therapy

N = 25,552

N = 24,116

N = 24,202

Variable

aOR

95% CI

aOR

95% CI

aOR

95% CI

  Missing  Child race or ethnicity   by home language   English    White    Black    Hispanic    Multiracial    Other   Non-English    White    Black    Hispanic    Multiracial    Other Family domain   Age position   Only child   Oldest child   Second oldest   Third oldest   Fourth oldest   Household income   ≤100% FPL   100%–199%   200%–399%   ≥ 400% FPL  Parent involvement (days   read to child per week)   0 days   1–3 days   4–7 days   No. adults in household   1   2   ≥3   Marital status   Married   Cohabiting   Neither   No parents Community domain  Provider ask about  concerns

0.57

[0.66, 1.94]

0.41

[0.06, 3.10]

6.52

[0.55, 77.60]  

Ref 1.2 1.02 1.07 1.19

Ref [0.98, 1.47] [0.78, 1.33] [0.81, 1.42] [0.85, 1.65]

Ref 0.50* 0.63 0.95 0.74

Ref [0.30, 0.81] [0.33, 1.20] [0.47, 1.90] [0.43, 1.27]

Ref 0.53* 0.89 0.29* 0.95

1.83 2.87* 1.59* 3.63* 3.22*

[0.84, 3.98] [1.59, 6.36] [1.20, 2.12] [1.02, 12.98] [1.89, 5.80]

0.48 [0.17, 1.36] N/A N/A 0.14* [0.05, 0.41] N/A N/A 0.07* [0.02, 0.26]

0.94 N/A 0.17* 2.71 0.07*

Ref [0.29, 0.94] [0.46, 1.70] [0.12, 0.74] [0.52, 1.74]     [0.35, 2.51] N/A [0.07, 0.40] [0.29, 25.40] [0.02, 0.33]

Ref 0.82* 0.80* 0.67* 0.60*

Ref [0.67, 1.00] [0.68, 0.95] [0.54, 0.84] [0.44, 0.82]

Ref 1.53 1.21 1.30 1.34

Ref [0.98, 2.39] [0.73, 2.01] [0.77, 2.21] [0.73, 2.47]

Ref 1.04 0.73 0.75 0.72

Ref 1.18 1.13 0.95

Ref [0.94, 1.48] [0.89, 1.44] [0.73, 1.24]

Ref 1.91* 1.73 1.50

Ref [1.08, 3.40] [0.94, 3.16] [0.70, 3.25]

Ref 1.68 2.46* 2.22*

Ref 1.49* 1.01

Ref [1.06, 2.10] [0.76, 1.41]

Ref 3.89* 4.60*

Ref [1.44, 10.52] [1.76, 12.00]

Ref 1.77 4.21*

Ref 0.93 0.77

Ref [0.70, 1.24] [0.58, 1.03]

Ref 0.81 1.45

Ref [0.38, 1.75] [0.73, 2.88]

Ref 1.53 2.18

Ref 1.15 1.12 0.77

Ref [0.88, 1.51] [0.84, 1.46] [0.34, 1.74]

Ref 0.96 0.88 2.82

[0.48, 1.91] [0.44, 1.74] [0.55, 14.36]

Ref 1.42 1.75 2.92

  Ref [0.60, 1.79] [0.49, 1.10] [0.36, 1.56] [0.35, 1.50]   Ref [0.94, 3.00] [1.31, 4.61] [1.05, 4.71]   Ref [0.62, 5.05] [1.44, 12.36]   Ref [0.69, 3.38] [0.96, 4.96]   Ref [0.71, 2.82] [0.86, 3.58] [0.41, 20.80]   (continued)

261

Marshall et al. Table 2. (continued)

 

Model 1: Concern

Model 2: IFSP/IEP

Model 3: Therapy

N = 25,552

N = 24,116

N = 24,202

Variable

aOR

95% CI

  Yes   No   Insurance type   Public   Private   None  Usual source of medical  care   Yes   No   Child care   Neither type   Nonrelatives   Relatives   Both types

1.19* Ref

[1.05, 1.36] Ref

Ref 0.88 0.79

aOR

95% CI

aOR

95% CI

2.09* Ref

[1.45, 3.00] Ref

2.17 Ref

Ref [0.73, 1.06] [0.60, 1.04]

Ref 0.44* 0.33*

Ref [0.25, 0.78] [0.17, 0.65]

Ref 0.55 0.25

[1.35, 3.49] Ref   Ref [0.32, 0.96] [0.11, 0.58]  

0.56* Ref

[0.40, 0.78] Ref

0.56 Ref

[0.25, 1.26] Ref

1.4 Ref

Ref 1.06 1.31* 1.12

Ref [0.91, 1.24] [1.07, 1.60] [0.90, 1.39]

Ref 1.16 0.64 1.19

Ref [0.83, 1.62] [0.37, 1.12] [0.73, 1.92

Ref 0.98 0.53 0.85

[0.73, 2.68] Ref   Ref [0.64, 1.52] [0.29, 0.95] [0.48, 1.51]

Note. IEP/IFSP = individualized education program or individualized family support plan; CI = confidence interval; FPL = federal poverty level; Ref = reference. N/A = statistics could not be calculated as one cell contains n = 0. *p ≤ .05.

English-speaking (home language) parents of Black children were half as likely (aOR = 0.50) as English-speaking parents of White children to have a child with an IFSP or IEP. The odds of IFSP or IEP enrollment were even lower for non-English speakers who were Hispanic (aOR = 0.14) or reported “other” race (aOR = 0.07), although non-English-speaking minorities were 1.59 to 3.63 times more likely to express concern about development in the univariate models. No statistically significant association was found with gross motor, academic, social, or behavior concerns or child care type. Child or maternal age, maternal education, marital status, or the number of adults in the home or siblings did not contribute increased or decreased likelihood but were controlled for in the model. Again, in the univariate models but not in the multivariable model, likelihood of an IFSP or IEP increased by child age (age >2, aOR = 2.11; age 3, aOR = 2.56; age 4, aOR = 4.08; age 5, aOR = 5.05); was statistically significantly higher among children with

younger siblings (aOR = 1.78), children who lived with neither parent (aOR = 2.68), and children in nonrelative care (aOR = 1.64) or both nonrelative and relative care (aOR = 1.99); and was lower among households with two adults (aOR = 0.64).

Private Therapy Services Parents of children who were male (aOR = 2.44) and had poorer reported health (for children in poorest health, aOR = 53.31), who read to the child most often (4–7 times per week, aOR = 4.21), and who had a provider ask about concerns (aOR = 2.17) were more likely to have children receiving private therapy. The likelihood of the child’s receiving private therapy increased with income above 200% FPL (aOR = 2.46 for 200%–399% FPL and 2.22 for >400% FPL) and maternal age (increasing with age up to aOR = 11.08 for mothers over 40). Children of mothers without a high school diploma were more likely to receive private therapy (aOR = 1.93).

262

Exceptional Children 82(2)

Table 3.  Summary of Significant Variables Across Three Outcomes: Concern, IEP/IFSP, and Therapy. Model 1: Concern Child domain   Child male (1.40)   Increasingly poorer child   health (1.81, 2.11, 5.88, 7.70) Not included in Model 1 Parent domain   Maternal education missing  (2.18)    

Model 2: Public Services (IEP/IFSP)

Model 3: Private Services (Therapy)

Child male (1.67) Child male (2.44) Relatively poorer child health Increasingly poorer child health (1.61, 2.10) (3.72, 8.53, 15.82, 53.31) General, speech, language, selfGeneral, speech, self-help, gross help, fine motor concern (4.18, motor concern (6.22, 3.38, 2.36, 4.63, 1.97, 1.63, 2.34) 3.10) Maternal education < HS (1.93) Older mother (7.00, 9.77, 11.08) English home language and Black English home language and Black race (0.50) or multirace (0.53, 0.29) Non-English home language Non-English home language (Hispanic, 0.14; other race, 0.07) (Hispanic, 0.17; other race, 0.07)

Non-English home language (Hispanic, 1.59; Black, 2.87; multiracial, 3.63) Family domain   Older siblings (0.60, 0.67, 0.80)   Income 100%–199%FPL (1.91)   Read to child 1–3 days vs.   none (1.49) Community domain   Provider ask about   concerns (1.19)  

  Higher income (200%–399% FPL, 2.46; >400% FPL, 2.22) Read to child more vs. none (1–3 Read to child most vs. none (4–7 days, 3.89; 4–7 days, 4.60) days, 5.32)

Provider ask about concerns (2.09) Private or no insurance (0.44, 0.33)

  Usual source of health   care (0.56)   Relative child care vs.   none (1.31)

Provider ask about concerns (2.17) Private or no insurance (0.55, 0.25)   Relative child care vs. none (0.53)

Note. Adjusted odds ratios shown in parentheses. Number of adults in the household and marital status were included in all models. IEP/IFSP = individualized education program or individualized family support plan; FPL = federal poverty level.

Families with private (aOR = 0.55) or no insurance (aOR = 0.25) were less likely to enroll children in private therapy compared to those with public insurance. English-speaking parents of Black children (aOR = 0.53) or children o multiple races (aOR = 0.29) were also less likely to report accessing private therapy services, as were non-English-speaking parents of Hispanic (aOR = 0.17) or “other” race or ethnicity (aOR = 0.07), even though these groups had higher odds of developmental concern. Finally, reports of “any” concern about learning, development, or behavior (aOR = 6.22) or specific concerns about speech (aOR = 3.38), self-help (aOR =

2.36), and gross motor skills (aOR = 3.10) were associated with increased odds of accessing therapy. No statistically significant association was found with other types of concerns. Child age, marital status, number of adults in the household or siblings, child care type, or usual source of medical care did not contribute increased or decreased likelihood of therapy enrollment but were controlled for in the model. In the univariate models but not in the multivariable model, likelihood of enrollment in therapy increased by child age (age >1, aOR = 2.24; age 2, aOR = 2.27; age 3, aOR = 2.93; age 4, aOR = 3.72; age 5, aOR = 4.12) and among families

Marshall et al. who read to their child 1 to 3 days (aOR = 3.88) or 4 to 7 days (aOR = 3.75) versus 0 days per week.

Discussion This study identified several child, parent, family, and community characteristics associated with increased or decreased likelihood of early recognition and enrollment in services for developmental delays among young children (summarized in Table 3). Although child factors had greater influence on parent concern, family- and community-level factors were also strongly related to use of services.

Likelihood of parent concern increased almost fivefold with child age from infancy to age 5 and sevenfold for children with the poorest reported health.

Developmental Concern Child factors. Likelihood of parent concern increased almost fivefold with child age from infancy to age 5 and sevenfold for children with the poorest reported health. These findings are intuitive; delays identified in early childhood often co-occur with health issues or birth defects and with developmental milestones for toddlers. The age trend may be also influenced by the increased focus on academic and social skills during prekindergarten years, the higher visibility of behavioral and speech and language issues (the most common concerns), and the fact that some developmental questions in the NSCH are not asked of parents with children under age 4 months, 10 months, or 18 months. The higher likelihood of developmental concern among parents of boys corroborates other research showing that boys are recognized with developmental problems more often, due to exhibiting more externalizing behaviors, developing some skills later than girls, or perhaps other social, cultural, or biological factors (Boyle et al., 2011; Else-Quest, Hyde, Goldsmith, & Van Hulle, 2006; Spelke, 2005). Parent and family factors.  The complex personal and social context in which developmental

263 concern among parents occurs is partially reflected in our findings. Increased concern among those reporting missing maternal education and among children living with no parents (in the bivariate analysis) could be related to some shared characteristic, such as children in foster or relative care who are at higher risk of developmental delays (Herman, 2007; Leiter & Rieker, 2012). English-speaking Black parents also had higher adjusted odds for concern, as did non-English-speaking, non-White parents. Disparities in child health and development by race and ethnicity (Stahmer, Sutton, Fox, & Leslie, 2008) may result from greater developmental risks associated with comparatively poorer birth outcomes among Black mothers (Curry, Pfeiffer, Slopen, & McVeigh, 2012; Rushton & Jensen, 2005) and also the differences in speech development among children who are second-language learners (Hambly, Wren, McLeod, & Roulstone, 2013; Hoff, Core, Place, Rumiche, & Parra, 2012). Racial disparities in early intervention were also found in other national longitudinal studies (Rosenberg et al., 2008; Walker et al., 2011). Levels of parent involvement and cognitive stimulation at home have been associated with increased and decreased risk of developmental delays (Walker et al., 2011). Parents who read to their child moderately reported higher odds of concern than those who did not read, yet those who read to the child most often did not have statistically significantly higher odds. The equivocal findings in our study may be due to its cross-sectional design; parents who read to the child provide more cognitive stimulation and thus lower risk of delay, and those who are reading to the child moderately may be working to ameliorate a developmental concern or may be more likely to notice an existing problem. Parental concern was less likely among children with older siblings and decreased as the number of siblings increased. Further research is needed to determine how parenting siblings or prior parenting experience could be related to risk, recognition, and concern for developmental delays. Community factors.  At the community level, the lower odds of concern among parents reporting a usual source of care may be due to the imprecise measure in the NSCH; a usual place of care

264 includes a doctor’s office, hospital outpatient department, clinic or health center, school, friend’s or relative’s home, “some other place,” or a telephone advice line but does not include a hospital emergency room, location outside the United States, or one place that is utilized inconsistently. Parents reporting developmental concerns accessed medical care for their child more often from the emergency room or a clinic than a personal doctor’s office, had public rather than private insurance, reported a larger number of doctor visits, and had children with poorer health than did parents without concerns. Parent recognition of delays may be more related to provider practices than to provider setting. The health care provider asking about concerns was strongly associated with increased recognition and services enrollment. Also, poorer health status may be associated with greater frequency of involvement with medical providers, including discussing concerns and establishing referrals. Further, providers working with higher-risk populations (e.g., low-income families) may be more likely to ask questions about developmental concerns (Anderson et al., 2003; Boyle et al., 2011).

Enrollment in Services Child, parent, and family factors.  Overall, enrollment in developmental services (public intervention or therapy) was strikingly low. In fact, fewer than 5% of parents reported their children were in either of the services, and a portion of those children (1.69% of the overall sample) was enrolled in both types of services simultaneously. Although child factors and health care provider practices influence parent concern and enrollment in services, parent action is also essential. Our finding that racialand ethnic-minority parents are less likely to access services despite being more likely to have concerns confirms other reports and could relate to issues with cultural competency and language barriers (Miller, Condin, McKellin, & Shaw, 2009; Paradis, 2010; Pinto-Martin et al., 2005; Rosenberg et al., 2008). Also, parents who read to their children more often were more likely to enroll them in public or private services, reflecting perhaps higher investment in child development (Davis-Kean,

Exceptional Children 82(2) 2005; Lee & Bowen, 2006). Mothers who were older and those with higher incomes were more likely to access therapy for their child. Perhaps these combined factors relate to self-efficacy, access to information, or access to community resources as described next. Although the reasons for these associations cannot be verified within the scope of this study, it is fair to say that parent characteristics appear to be important in some way to the process of successfully enrolling children in developmental services. Community factors. Barriers to accessing services cited in the literature include cost and lack of insurance coverage (Pinto-Martin et al., 2005). Although we also identified these barriers, the relationship between income and insurance and access to developmental services is not necessarily linear. Families with lower income may receive developmental screening via Medicaid’s mandated Early and Periodic Screening, Diagnosis, and Treatment schedule and may participate in programs focused on “at-risk” populations, such as subsidized child care or Head Start, that implement developmental screening and referral. Conversely, higher-income families may have access to high-quality pediatric care and child care, increasing opportunities for developmental screening, consultation, and referral to specialists. These hypotheses were not confirmed in the study, likely because the rates of screening were low overall. Parents who had higher income had greater odds of child enrollment in therapy; those enrolled in public intervention had income that was above Medicaid eligibility but below 200% FPL. Public insurance, versus private or no insurance, was associated with increased likelihood of services, reflecting funding barriers to accessing therapy through private insurance or private pay, the important role of Medicaid in funding screening and intervention services, and the additional access to public intervention that programs serving low-income families (such as Head Start or subsidized child care) facilitates. Finally, some state intervention programs may partially cover therapy services whereas others may fully cover therapy, regardless of the family’s income or insurance coverage.

Marshall et al. As the concern variable was derived from questions about specific types of concerns, these variables could not be examined relative to parent concern (recognition of delay); however, some specific types of concern (speech, receptive language, fine motor) were statistically significantly associated with higher likelihood of enrollment in public intervention and special education, and others (speech, self-help, gross motor) were associated with higher likelihood of enrollment in private therapy. These differences may reflect priorities and resources available in public schools or referral to private therapy by physicians concerned with health-related issues, such as feeding disorders or motor impairment. Most certainly, the health care and child care providers must be aware of available programs, services, and providers in order to make referrals for all children who need them. Younger children with subtle cognitive delays or social and behavioral challenges may be less likely to be referred or to qualify for services.

Study Limitations and Strengths One major limitation in this study was its cross-sectional design; we were unable to ascertain which associations may be related to risk factors for developmental delay in children versus factors facilitating parental recognition of developmental issues. Also, SAS 9.3 PROC SURVEY commands calculate descriptive and inferential statistics while accounting for categorical classification variables and sampling yet do not allow for multilevel modeling or hierarchical analyses. There are also limitations with using the NSCH data set: Children who have developmental delays but have not activated parent concern nor received services would not be identified. The analysis is also limited to those questions asked within the NSCH protocol. For example, the first question related to parent concern in general asked whether the parent had “any concern” related to the child’s learning, behavior, or development. Although only 9% answered in the affirmative, 38% of parents reported a concern when asked additional probes about specific skills. Parents may perceive or express concerns differently depending on how

265 a question is worded; therefore a composite variable was created that included the variable capturing any concern as well as the variables related to specific concerns in any of the developmental domains. Although the presence of a developmental delay is not confirmed in this study by a standardized screening tool, research has suggested that to some extent parent concern is a reliable indicator of a delay (Glascoe, 1997). However, the level of concern may be subjective, and there is no other way to distinguish the severity of the child’s delay. It is also important to consider that parents may report that their child has an IFSP or IEP or is enrolled in therapy yet may not distinguish between the two, as public intervention may include therapy. Our analysis is based on the assumption that if the child is receiving therapy, the therapist is providing a service that the child needs. Also, the race/ethnicity variable in the NSCH is derived from a question concerning the selected child’s race, not the race or ethnicity of the parent, and the primary-language question refers to language spoken in the household. The NSCH also does not include measures of parenting self-efficacy, knowledge of child development, or qualities that may influence recognition or help-seeking; measures used in our study are limited to parenting experience (siblings), maternal education, and number of days read to the child (which showed statistically significant associations across all three outcomes, as has been found to be important in other studies; Davis-Kean, 2005). Social support was limited to marital status, number of adults in the household, and relative child care; therefore the influence of trusted others warrants further investigation. There also was no direct measure of help-seeking in the NSCH, thus enrollment in services was examined. However, the distinction between those who enroll in public versus private services is enlightening, as it points out potential facilitators and barriers to service options for a variety of children and families. Study strengths include its nationally representative sample, the inclusion of several covariates in multiple socioecological domains in the analysis, and its breadth of parents with and without concerns who may or may not have enrolled in services. Our results can guide efforts to improve early recognition of developmental

266 delays and the timeliness of early intervention. For example, child-level data can be used as a barometer for where targeted screening and outreach does and should occur to improve early identification and enrollment in services. System-level barriers, such as enrollment categories, cost and insurance coverage, and racial and ethnic disparities can easily be identified and efforts made to address them. Finally, future research must strongly consider parent factors influencing the process—there was some evidence in the study that parent factors play an important role.

Conclusion This study provided insights into factors that may be related to recognition of developmental delays and differences in access to public and private services (Table 3). Although it is important to identify children at risk for delay— particularly at the earliest ages and among those with substantial health concerns—this study also demonstrates that delays and concerns span all ages, income levels, races, ethnicities, and family structures. Perhaps children of the nearly 40% of parents who had concerns about their behavior or development will not all require formal intervention services, but as pointed out by Rosenberg et al. (2013), depending on the definitions of developmental delay or high risk of developmental delay, up to 40% of children could require some level of developmental support. Also, in planning intervention services funding and outreach, policy makers must recognize the differences in awareness and referral for specific types of concerns and particularly the large gap in support for behavioral concerns, which are among the most common concerns. Although the examination of help-seeking was limited to those enrolled in services, many parents have concerns and face potential gaps in access to services, especially parents who are racial, ethnic, or linguistic minorities. Insurance and financial gaps remain barriers to gaining access to service options, as do cultural differences, particularly among families whose home language is not English. First, it is important to examine patterns of participa-

Exceptional Children 82(2) tion in early intervention services by race and ethnicity because of prior research highlighting underutilization of mental health or developmental services (Bussing, Zima, Faye, & Garvan, 2003; Featherstone & Broadhurst, 2003) and disproportionate underrepresentation in early intervention and overrepresentation in special education (U.S. Department of Education, 2013) and school suspensions and expulsions (Fenning & Rose, 2007) for racial or ethnic minority children. Systems of early identification, referral, and intervention must continue to attend to cultural sensitivity and the dearth of information and support for families with limited English proficiency. Second, family income (corresponding with maternal age and education) often determines the type and extent of insurance coverage for primary health care and also dictates access to the type and quality of health care, child care, and other services that bear responsibility for early identification of developmental delays, response to parents’ developmental concerns, and referral. Finally, in addition to continuing to address system-level barriers to services access, parents must know where to go for information when developmental questions or concerns arise; a multitude of strategies to increase awareness of development milestones, red flags, and intervention services should continue to be utilized populationwide, among parents and professionals. Future study needs to identify the important role of parenting practices related to monitoring and supporting the child’s development (reading to the child was an important variable in this study). The role of these additional factors—access to information and family culture, dynamics, and involvement—should be explored in order to improve uptake of social marketing campaigns as well as follow-through on concerns and referrals. Also critical is the role of trusted others in the parents’ social network in facilitating recognition and access to services, and particularly crucial is the role of health care providers in eliciting parent concerns and providing referrals to services. In this sample of over 27,000 parents, fewer than half (46%) reported that their primary health care provider asked

Marshall et al. them about developmental concerns. Child care providers may also offer developmental screening and referral; however, it was not consistent enough to show as a statistically significant influence on the process in this national study. Further examination of the timing and process for seeking early intervention requires longitudinal, prospective, populationwide studies, which could best be captured by databases of screening, referral, services, and child outcomes. Meanwhile, the results of this study can contribute to the body of information that helps to inform and ultimately improve community education, developmental screening, and child-find outreach efforts to enhance early identification and intervention for developmental delays. References Anderson, L. M., Shinn, C., Fullilove, M. T., Scrimshaw, S. C., Fielding, J. E., Normand, J., & Carande-Kulis, V. G. (2003). The effectiveness of early childhood development programs: A systematic review. American Journal of Preventive Medicine, 24(3), 32–46. doi:10.1016/S0749-3797(02)00655-4 Blumberg, S. J., Foster, E. B., Frasier, A. M., Satorius, J., Skalland, B. J., Nysse-Carris, K. L., . . . O’Connor, K. S. (2012). Design and operation of the National Survey of Children’s Health, 2007. Vital and Health Statistics: Ser. 1, Programs and Collection Procedures, 55, 1–149. Boyle, C. A., Boulet, S., Schieve, L. A., Cohen, R. A., Blumberg, S., Yeargin-Allsopp, M., . . . Kogan, M. D. (2011). Trends in the prevalence of developmental disabilities in US children, 1997–2008. Pediatrics, 127, 1034–1042. doi:10.1542/peds.2010-2989 Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press. Bussing, R., Zima, B. T., Gary, F. A., & Garvan, C. W. (2003). Barriers to detection, help-seeking, and service use for children with ADHD symptoms. Journal of Behavioral Health Services and Research, 30, 176–189. Children’s Defense Fund. (2014) State of America’s Children. Retrieved from http:// www.childrensdefense.org/library/stateof-americas-children/2014-soac.pdf?utm_

267 source=2014-SOAC-PDF&utm_medium=link &utm_campaign=2014-SOAC Chung, C., Liu, W., Chang, C., Chen, C., Tang, S. F., & Wong, A. M. (2011). The relationship between parental concerns and final diagnosis in children with developmental delay. Journal of Child Neurology, 26, 413–419. doi:10.1177/0883073810381922 Curry, A. E., Pfeiffer, M. R., Slopen, M. E., & McVeigh, K. H. (2012). Rates of early intervention referral and significant developmental delay, by birthweight and gestational age. Maternal and Child Health Journal, 16, 989– 996. doi:10.1007/s10995-011-0820-y Davis-Kean, P. E. (2005). The influence of parent education and family income on child achievement: the indirect role of parental expectations and the home environment. Journal of Family Psychology, 19, 294–304. doi:10.1037/08933200.19.2.294 Else-Quest, N. M., Hyde, J. S., Goldsmith, H. H., & Van Hulle, C. A. (2006). Gender differences in temperament: A meta-analysis. Psychological Bulletin, 132, 33–72. doi:10.1037/00332909.132.1.33 Featherstone, B., & Broadhurst, K. (2003). Engaging parents and carers with family support services: What can be learned from research on help-seeking? Child and Family Social Work, 8, 341–350. Fenning, P., & Rose, J. (2007). Overrepresentation of African American students in exclusionary discipline the role of school policy. Urban Education, 42, 536–559. Glascoe, F. P. (1997). Parents’ concerns about children’s development: Prescreening technique or screening test? Pediatrics, 99, 522–528. doi:10.154/peds.99.4.522 Glascoe, F. P. (2013). Parents’ Evaluation of Developmental Status (PEDS): Collaborating with parents. Nolensville, TN: PEDSTest.com. Guevara, J. P., Gerdes, M., Localio, R., Huang, Y. V., Pinto-Martin, J., Minkovitz, C. S., . . . Pati, S. (2013). Effectiveness of developmental screening in an urban setting. Pediatrics, 131, 30–37. doi:10.1542/peds.2012-0765 Hambly, H., Wren, Y., McLeod, S., & Roulstone, S. (2013). The influence of bilingualism on speech production: a systematic review. International Journal of Language and Communication Disorders, 48, 1–24. Herman, B. (2007). CAPTA and early childhood intervention: Policy and the role of parents (CCC Code 1532-8759/07). Washington, DC: National Association of Social Workers.

268 Hix-Small, H., Marks, K., Squires, J., & Nickel, R. (2007). Impact of implementing developmental screening at 12 and 24 months in a pediatric practice. Pediatrics, 120, 381–389. Hoff, E., Core, C., Place, S., Rumiche, R., & Parra, M. (2012). Dual language exposure and early bilingual development. Journal of Child Language, 39, 1–27. doi:10.1017/ S0305000910000759 Lee, J. S., & Bowen, N. K. (2006). Parent involvement, cultural capital, and the achievement gap among elementary school children. American Educational Research Journal, 43, 193–218. Leiter, V., & Rieker, P. P. (2012). Mind the gap: Gender differences in child special health care needs. Maternal and Child Health Journal, 16, 1072–1080. doi:10.1007/s10995-011-0834-5. Lu, M. C., & Halfon, M. (2003). Racial and ethnic disparities in birth outcomes: A life-course perspective. Maternal and Child Health Journal, 7, 13–30. doi:10.1023/A:1022537516969 Miller, A. R., Condin, C. J., McKellin, W. H., & Shaw, N. (2009). Continuity of care for children with complex chronic health conditions: parents’ perspectives. BMC Health Services Research, 9, 242. doi:10.1186/1472-6963-9-242 Paradis, J. (2010). The interface between bilingual development and specific language impairment. Applied Psycholinguistics, 31, 227–252. doi:10.1017/S0142716409990373 Pinto-Martin, J. A., Dunkle, M., Earls, M., Fliedner, D., & Landes, C. (2005). Developmental stages of developmental screening: steps to implementation of a successful program. American Journal of Public Health, 95, 1928–1932. Ramey, C. T., & Ramey, S. L. (2004). Early learning and school readiness: Can early intervention make a difference? Merrill-Palmer Quarterly, 50, 471–491. doi:10.1353/mpq.2004.0034 Rosenberg, S. A., Robinson, C. C., Shaw, E. F., & Ellison, M. C. (2013). Part C early intervention for infants and toddlers: Percentage eligible versus served. Pediatrics, 131, 38–46. Rosenberg, S. A., Zhang, D., & Robinson, C. C. (2008). Prevalence of developmental delays and participation in early intervention for young children. Pediatrics, 121, e1503–e1509. doi:10.1542/peds.2007-1680 Roux, A. M., Herrera, P., Wold, C. M., Dunkle, M. C., Glascoe, F. P., & Shattuck, P. T. (2012). Developmental and autism screening through 2-1-1: Reaching underserved families. American Journal of Preventive Medicine, 43(6S5), S457–S463. doi:10.1016/j. amepre.2012.08.011

Exceptional Children 82(2) Rushton, J. P., & Jensen, A. R. (2005). Thirty years of research on race differences in cognitive ability. Psychology, Public Policy, and Law, 11, 235–294. doi:10.1037/1076-8971.11.2.235 Seligman, M., & Darling, R. B. (2007). Ordinary families, special children: A systems approach to childhood disability (3rd ed.). New York, NY: Guilford Press. Shevell, M. (2008) Global developmental delay and mental retardation or intellectual disability: Conceptualization, evaluation, and etiology. Developmental Disabilities, Part I, 55, 1071–1084. Smith, L. E., Akai, C. E., Klerman, L. V., & Keltner, B. R. (2010). What mothers don’t know and doctors don’t say: Detecting early developmental delays. Infant Mental Health Journal, 31, 455–466. doi:10.1002/imhj.20266 Spelke, E. S. (2005). Sex differences in intrinsic aptitude for mathematics and science? A critical review. American Psychologist, 60, 950– 958. doi:10.1037/0003-066X.60.9.95 Stahmer, A. C., Sutton, D. T., Fox, L., & Leslie, L. K. (2008). State Part C agency practices and the Child Abuse Prevention and Treatment Act (CAPTA). Topics in Early Childhood Special Education, 28, 99–108. U.S. Department of Education. (2013), Thirtyfifth annual report to Congress on the implementation of the Individuals with Disabilities Education Act, Parts B and C. Retrieved from http://www2.ed.gov/about/reports/annual/ osep/2013/parts-b-c/index.html Valicenti-McDermott, M., Hottinger, K., Seijo, R., & Shulman, L. (2012). Age at diagnosis of autism spectrum disorders. Journal of Pediatrics, 161, 554–556. doi:10.1016/j. jpeds.2012.05.012 Walker, S. P., Wachs, T. D., Grantham-McGregor, S., Black, M. M., Nelson, C. A., Huffman, S. L., . . . Richter, L.(2011). Inequality in early childhood: Risk and protective factors for early child development. Lancet, 378, 1325– 1338. doi:10.1016/S0140-6736(11)60555-2

Authors’ Note We extend our thanks to Donna J. Petersen for her careful review and thoughtful appraisal of this project. The authors would also like to acknowledge the Child and Adolescent Health Measurement Initiative Data Resource Center for making these data available. Manuscript received June 2014; accepted March 2015.