Birth characteristics of children who used early intervention and special education services in New York City

Abstract

Background

Early intervention (EI) and special education (SE) are beneficial for children with developmental disabilities and/or delays and their families, yet there are disparities in service use. We sought to identify the birth characteristics that predict EI/SE service use patterns.

Methods

We conducted a retrospective cohort study using linked administrative data from five sources for all children born in 1998 to New York City resident mothers. Multinomial regression was used to identify birth characteristics that predicted predominant patterns of service use.

Results

Children with service use patterns characterized by late or limited/no EI use were more likely to be first-born children and have Black or Latina mothers. Children born with a gestational age ≤31 weeks were more likely to enter services early. Early term gestational age was associated with patterns of service use common to children with pervasive developmental delay, and maternal obesity was associated with the initiation of speech therapy at the time of entry into school.

Conclusions

Maternal racial disparities existed for patterns of EI/SE service use. Specific birth characteristics, such as parity and gestational age, may be useful to better identify children who are at risk for suboptimal EI use.

Introduction

Under Part B of the Individuals with Disabilities Education Act (IDEA), the United States Congress mandates public special education (SE) services, which includes early childhood special education (ECSE) and SE for kindergarten through grade 12 (K-12), for children with developmental disabilities or delays.¹ An Individualized Educational Program (IEP) is developed for each child eligible for ECSE/K-12 SE and includes individualized annual academic, social, behavioral and/or physical goals and how to achieve them through SE and related services, including occupational therapy (OT), physical therapy (PT), and speech therapy (ST).² For non-school-age children (i.e. infants and toddlers), IDEA Part C authorizes federal funding for states to administer early intervention (EI) services. As part of EI, an Individualized Family Service Plan (IFSP) is developed for each eligible child and their family that lays out their goals and what services (e.g. OT, PT, ST and special instruction [SI]) they should receive to achieve them.³

The use of IDEA services may vary across the spectrum of EI, ECSE and SE. For example, some children might receive EI from infancy and require ECSE/SE as they grow older. Other children might be referred to EI later for condition-specific reasons, such as autism spectrum disorder (ASD), which cannot be diagnosed as early as some other conditions.⁴ Children might cease the use of services after EI for a variety of reasons, most commonly because services are no longer needed,⁵ whereas other children might only access ECSE/SE because some conditions, such as learning disabilities, might not be apparent until children grow older.⁶

Despite the wide inclusivity of children with delays or disabilities, there are disparities in EI and SE service use related to race, ethnicity and nativity. For example, compared with White children, Black and Asian children are less likely to be enrolled in services, as are children of foreign-born and non-English-speaking mothers compared with children of US-born and English-speaking mothers.^7–11 Even among children who are enrolled, there are disparities with regard to intensity of services. For example, Black, low-income, and publicly insured children receive less intensive EI services than their White counterparts with fewer social risk factors.¹² These disparities are troubling, as EI has a positive effect on motor and cognitive outcomes for children and has psychosocial benefits for families^13–17; furthermore, SE later in childhood has a positive effect on learning-related behaviors.¹⁸ Therefore, ensuring timely and adequate use of EI and SE for children with developmental delays or disabilities should be a public health priority.

Children who enter EI late or with limited service use but then go on to require intensive services in the school setting are possible groups to target to optimize service use. Therefore, to aid in the early identification of these children, we sought to describe the birth characteristics of children with different patterns of EI and SE service use in New York City (NYC) and identify potential predictors of these patterns.

Methods

We conducted a retrospective cohort study using linked administrative data from the Longitudinal Study of Early Development (LSED) data warehouse. For this analysis, we used individual-level data for all children born in NYC in 1998 to resident mothers and who remained alive through third grade. LSED used a probabilistic approach to link children and siblings across four data sources from the NYC Department of Health and Mental Hygiene (EI Program, Lead Poisoning Prevention Program, Birth Certificate Registry, Death Certificate Registry) and the NYC Department of Education (DOE) SE and testing databases.¹⁹ This analysis was approved by the New York City Department of Health and Mental Hygiene Institutional Review Board.

Our primary outcome of interest was EI and SE service use pattern, based on our previous work that identified five predominant non-overlapping service use patterns from birth through third grade: (i) multiple therapies across EI/ECSE/SE; (ii) EI without transition to DOE schools or services; (iii) EI and intermittent ECSE/SE; (iv) older entry to EI and both ST and OT throughout ECSE/SE; and (v) limited EI use and mostly ST in ECSE/SE.²⁰

Birth characteristics included measures of maternal demographics, pregnancy risk factors, and birth outcomes, all captured in the aforementioned administrative sources:

Maternal demographics: race, nativity, country of origin if foreign-born, education level, employment status during pregnancy, borough of usual residence at birth, neighborhood poverty of residence at birth (percentage of population in the mother’s ZIP code living below 100% of the federal poverty level per the American Community Survey), participation in any public assistance program, and documentation of paternity.

Pregnancy risk factors: maternal age, paternal age, having no previous children, inadequate prenatal care (no prenatal care, inadequate or intermediate versus adequate or better²¹), any tobacco use during pregnancy, any alcohol or illicit drug use during pregnancy, any maternal medical risk factor (Appendix 1), obesity (defined as ≥200 pounds prior to pregnancy²²), maternal hypertension (including pre-existing hypertension, pre-eclampsia, eclampsia), maternal diabetes (including pre-existing diabetes and gestational diabetes), any labor or delivery complications (Appendix 1), and if Medicaid was a payer at the time of delivery.

Birth outcomes: child sex, being part of a multiple gestation, gestational age, birth weight, small for gestational age < 10th percentile, required neonatal intensive care unit (NICU) at birth, 5-minute Apgar score <7, any abnormal condition as a newborn, possible neonatal hypoxic condition, and any congenital anomaly (see Appendix 1 for the last three variables).

For children born in 1998 to NYC resident mothers who used EI/SE services, we computed frequencies for selected birth characteristics overall and stratified by the five predominant service use patterns. Next we conducted multinomial logistic regression to determine predictors of each pattern using pattern 3 as a reference group because these children only had intermittent use of SE after EI and had third grade test scores similar to citywide means.^23,24 All analyses were done in SAS 9.4 (SAS Institute, Inc., Cary, NC) and a P value below 0.05 was considered statistically significant.

Results

In 1998, 113 627 children were born to NYC resident mothers and did not have a death record (N = 574); 11.3% (N = 12 806) had some indication of IDEA service use across EI, ECSE, and/or SE; 77.8% (N = 88 420) had no indication of service use; and 10.9% (N = 12 401) had no record after birth.

Multiple therapies across EI/ECSE/SE (pattern 1)

Pattern 1 made up 13.3% of children with IDEA service use. For maternal demographic characteristics (Table 1), compared with the overall population of children who used services, children in pattern 1 more often had a Black mother (30.3 versus 26.7%). For pregnancy risk factors (Table 2), pattern 1 children more often had a mother ≥35 years (24.4 versus 18.8%), a mother with any medical risk factor (33.6 versus 29.6%), and a mother with any labor or delivery complication (44.9 versus 40.0%). For birth outcomes (Table 3), children in this pattern were more often part of a multiple gestation (8.2 versus 6.4%), small for gestational age (22.6 versus 17.6%), required NICU care (28.2 versus 17.3%), had a 5-minute Apgar score <7 (5.7 versus 2.3%), had an abnormal condition as a newborn (12.5 versus 7.4%), had a possible neonatal hypoxic neonatal condition (11.1 versus 7.8%) and had a congenital anomaly (4.2 versus 1.6%).

Table 1.

Demographic characteristics of children born in 1998 in New York City to resident mothers by early intervention and special education service use pattern

	Any EI/SE service use	Multiple therapies across EI/ECSE/SE (pattern 1)	EI without transition to DOE schools or services (pattern 2)	EI and intermittent ECSE/SE (pattern 3)	Older entry to EI and both ST & OT throughout ECSE/SE (pattern 4)	Limited EI use and mostly ST in ECSE/SE (pattern 5)
N	12 806	1698	2969	2271	1098	4770
Maternal race, N = 113 296
Black	26.7	30.3	24.8	25.4	32.6	25.9
White	28.7	27.1	39.2	29.9	23.5	23.3
Latina	39.1	36.7	31.0	39.1	39.5	45.0
Asian/PI	5.3	5.7	4.8	5.4	4.1	5.6
Other	0.2	0.1	0.2	0.2	0.3	0.1
Maternal nativity, N = 112 950
Foreign born	44.7	43.0	39.8	47.4	40.9	47.9
Maternal origin if foreign born, N = 56 704
Canada	0.7	0.4	1.4	0.7	0.7	0.4
Mexico/Central America/Caribbean	56.2	54.1	46.0	54.6	60.5	62.0
South America	14.4	16.3	14.3	14.9	14.4	13.6
Europe	8.8	8.5	13.6	11.2	4.5	6.1
Africa	4.8	4.8	5.4	3.8	7.6	4.3
Asia and Near East	15.3	15.8	19.3	14.9	12.3	13.8
Maternal education, N = 111 254
< 12 years	29.9	28.2	22.6	29.9	30.2	34.9
12 years	35.6	38.2	34.1	36.3	32.3	35.9
> 12 years	34.6	33.7	43.3	33.8	37.5	29.2
Maternal employment during pregnancy, N = 113 627
Yes	30.0	30.6	35.1	29.5	30.5	26.6
Borough of mother’s usual residence at birth, N = 113 627
Manhattan	15.6	14.0	17.1	14.6	16.5	15.4
Bronx	19.9	20.0	15.8	20.4	23.1	21.5
Brooklyn	36.0	36.9	39.6	37.1	29.0	34.7
Queens	22.4	24.3	21.2	21.9	23.6	22.4
Staten Island	6.1	4.7	6.4	6.0	7.8	6.1
Neighborhood poverty, N = 113 474
< 10% (very low)	16.8	16.6	20.8	16.2	16.4	14.8
10–19%	23.2	23.1	24.7	23.1	23.9	22.2
20–29%	24.2	24.6	23.5	24.4	23.8	24.4
≥30% (very high)	35.8	35.8	30.9	36.3	36.0	38.6
Maternal participation in any public assistance program(s), N = 113 627
Yes	49.7	49.1	40.5	51.5	49.8	54.7
Documentation of paternity, N = 113 627
Marriage certificate	49.5	49.5	57.0	50.1	47.5	45.0
Affidavit of paternity	27.7	26.5	22.1	27.0	30.1	31.3
Neither	22.9	24	21.0	22.9	22.4	23.7

Column percentages for each group are presented. Columns may not add up to 100 due to rounding

Table 2.

Pregnancy risk factors of children born in 1998 in New York City to resident mothers by early intervention and special education service use pattern

	Any EI/SE service use	Multiple therapies across EI/ECSE/SE (pattern 1)	EI without transition to DOE schools or services (pattern 2)	EI and intermittent ECSE/SE (pattern 3)	Older entry to EI and both ST and OT throughout ECSE/SE (pattern 4)	Limited EI use and mostly ST in ECSE/SE (pattern 5)
N	12 806	1698	2969	2271	1098	4770
Maternal Age, N = 113 620
<25 years	31.9	26.7	28.0	30.6	30.0	37.3
25–34 years	49.3	48.9	52.3	51.1	49.5	46.7
≥35 years	18.8	24.4	19.7	18.3	20.5	16.0
Paternal age, N = 113 627
<35 years	66.5	60.6	65.4	65.1	65.9	70.3
35–44 years	28.0	32.3	28.9	29.6	28.5	25.0
≥45 years	5.5	7.1	5.7	5.2	6.7	4.7
No previous children, N = 113 627
Yes	39.4	37.8	39.0	36.2	44.7	40.6
Inadequate prenatal care, N = 113 627
Yes	40.1	39.5	38.6	40.3	36.3	41.9
Mother had any tobacco use during pregnancy, N = 113 453
Yes	6.1	7.4	6.2	6.6	4.7	5.5
Mother had any alcohol or drug use during pregnancy, N = 113 453
Yes	2.9	3.3	3.5	3.3	2.8	2.2
Mother had any medical risk factor, N = 113 627
Yes	29.6	33.6	32.1	29.3	30.7	26.6
Maternal obesity, N = 113 627
Yes	8.3	9.2	8.1	7.4	7.4	8.8
Maternal diabetes, N = 113 627
Yes	4.0	6.2	4.7	4.6	4.8	5.4
Maternal hypertension, N = 113 627
Yes	6.1	7.1	6.4	6.6	6.8	5.2
Any labor or delivery complications, N = 113 627
Yes	40.0	44.9	41.2	39.3	41.4	37.6
Medicaid was payer at time of delivery, N = 113 627
Yes	56.3	54.4	46.8	57.4	55.7	62.5

Column percentages for each group are presented. Columns may not add up to 100 due to rounding

Table 3.

Birth outcomes of children born in 1998 in New York City to resident mothers by early intervention and special education service use pattern

	Any EI/SE service use	Multiple therapies across EI/ECSE/SE (pattern 1)	EI without transition to DOE schools or services (pattern 2)	EI and intermittent ECSE/SE (pattern 3)	Older entry to EI and both ST and OT throughout ECSE/SE (pattern 4)	Limited EI use and mostly ST in ECSE/SE (pattern 5)
N	12 806	1698	2969	2271	1098	4770
Child sex, N = 113 627
Male	66.5	65.7	63.3	62.8	82.3	67.0
Child was part of multiple gestation, N = 113 627
Yes	6.4	8.2	9.3	6.7	4.6	4.1
Gestational age, N = 112 874
≤31 weeks	6.1	13.5	9.7	5.8	3.6	2.0
32–33 weeks	2.6	3.5	3.8	2.6	2.2	1.7
34–36 weeks	9.3	11.5	10.5	9.5	8.9	7.7
37–38 weeks	22.0	21.3	20.9	21.5	26.0	22.3
39–40 weeks	50.2	42.3	46.0	51.4	48.8	55.3
≥41 weeks	9.8	8.0	9.3	9.2	10.5	11.0
Birth weight, N = 113 598
<1000 g	3.1	2.7	7.9	2.7	1.6	0.6
1000–1499 g	2.9	3.3	5.4	3.3	1.3	1.3
1500–2499 g	11.6	11.3	16.1	11.3	9.5	8.8
2500–3999 g	74.4	74.2	63.5	74.2	78.0	80.8
≥4000 g	8.0	8.5	7.1	8.5	9.6	8.6
Small for gestational age, N = 112 878
Yes	17.6	16.1	22.6	16.1	15.7	15.4
Child required NICU, N = 113 571
Yes	17.3	17.3	28.2	17.3	13.8	11.6
Child’s 5-minute Apgar score, N = 113 627
< 7	2.3	1.9	5.7	1.9	1.3	1.0
Child had any abnormal condition as a newborn, N = 113 627
Yes	7.4	7.0	12.5	7.0	5.7	4.9
Child had possible neonatal hypoxic condition, N = 113 627
Yes	7.8	11.1	8.8	7.2	7.6	6.4
Child had any congenital anomaly, N = 113 627
Yes	1.6	4.2	2.0	1.1	1.3	0.8

Column percentages for each group are presented. Columns may not add up to 100 due to rounding

EI without transition to DOE schools or services (pattern 2)

Pattern 2 made up 23.2% of children who used services. For maternal demographic characteristics (Table 1), compared with the overall population of children who used services, children in pattern 3 more often had a White mother (39.2 versus 28.7%) and less often a foreign-born mother (39.8 versus 47.4%). If they did have a foreign-born mother, they more often had one from Asia or the Near East (19.3 versus 15.3%). Children in this pattern less often had a mother with < 12 years of education (22.6 versus 29.9%), more often had a mother who was employed during the pregnancy (35.1 versus 30.0%), less often lived in very high poverty neighborhoods (30.9 versus 35.8%), and less often had a mother participating in any public assistance program (40.5 versus 49.7%). Children in this pattern also more often had a marriage certificate used as documentation of paternity (57.0 versus 49.5%). This pattern had a similar distribution of pregnancy risk factors (Table 2) compared with the overall population of children who used services, except for Medicaid being a payer at the time of delivery, which was less common (46.8 versus 56.3%). For birth outcomes (Table 3), children in this pattern were more often part of a multiple gestation (9.3 versus 6.4%), small for gestational age (20.1 versus 17.6%), and required NICU care (21.6 versus 17.3%).

EI and intermittent ECSE/SE (pattern 3)

Pattern 3 made up 17.7% of children who used services. Children in this pattern had a similar distribution of all maternal demographic characteristics (Table 1), pregnancy risk factors (Table 2), and birth outcomes (Table 3) relative to the overall population of children who used services.

Older entry to EI and both ST & OT throughout ECSE/SE (pattern 4)

Pattern 4 made up 8.6% of children with IDEA service use. For maternal demographic characteristics (Table 1), compared with the overall population of children who used services, children in pattern 4 more often had a Black mother (32.6 versus 26.7%). For pregnancy risk factors (Table 2), children in this pattern more often had a mother without previous children (44.7 versus 39.4%), less often had inadequate prenatal care (36.3 versus 40.1%), and less often had a mother who used tobacco during pregnancy (4.7 versus 6.1%). For birth outcomes (Table 3), children in this pattern were more often male (82.3 versus 66.5%) and less often had a gestational age ≤31 weeks (3.6 versus 6.1%), less often required NICU care (13.8 versus 17.3%), less often had a 5-minute Apgar score <7 (1.3 versus 2.3%) and less often had an abnormal condition as a newborn (5.7 versus 7.4%).

Limited EI use and mostly ST in ECSE/SE (pattern 5)

Pattern 5 made up 37.3% of children who used services. For maternal demographic characteristics (Table 1), compared with the overall population of children who used services, children in pattern 5 more often had a Latina mother (45.0 versus 39.1%), more often had a foreign-born mother (47.9 versus 44.7%) and, if foreign-born, more often had a mother who came from Mexico/Central America/Caribbean (62.0 versus 56.2%). Children in this pattern more often had a mother with < 12 years of education (34.9 versus 22.9%), less often had a mother who was employed during the pregnancy (26.6 versus 30.0%) and more often had a mother participating in any public assistance program (54.7 versus 49.7%; Table 1). For pregnancy risk factors (Table 2), children in this pattern more often had a mother < 25 years old (37.3 versus 31.9%) and had Medicaid as a payer at the time of delivery (62.5 versus 56.3%). For birth outcomes (Table 3), children in this pattern less often were part of a multiple gestation (4.1 versus 6.4%), had a gestational age ≤31 weeks (2.0 versus 6.1%), less often required NICU care (11.6 versus 17.3%), had a 5-minute Apgar score <7 (1.0 versus 2.3%), had any abnormal condition as a newborn (4.9 versus 7.4%), and had a congenital anomaly (0.8 versus 1.6%).

Predictors of service use pattern

In multivariable multinomial logistic regression using pattern 3 as a reference group, specific predictors of each service use pattern emerged (Table 4).

Table 4.

Multinomial regression model identifying predictors of service use pattern comparing with children who received early EI and only intermittent ECSE/SE (pattern 3)

	Multiple therapies across EI/ECSE/SE (versus pattern 3) aOR (95% CI)	EI without transition to DOE schools or services (versus pattern 3) aOR (95% CI)	Older entry to EI and both ST and OT throughout ECSE/SE (versus pattern 3) aOR (95% CI)	Limited EI use and mostly ST in ECSE/SE (versus pattern 3) aOR (95% CI)
Maternal race
Black	1.27 (1.03–1.57)	0.87 (0.73–1.05)	2.05 (1.60–2.62)	1.41 (1.19–1.68)
White	Reference	Reference	Reference	Reference
Latina	1.25 (1.01–1.54)	0.84 (0.70–1.00)	1.49 (1.16–1.91)	1.43 (1.21–1.69)
Asian/PI	1.27 (0.91–1.75)	0.80 (0.60–1.07)	1.08 (0.72–1.63)	1.38 (1.07–1.79)
Other	0.98 (0.16–6.02)	1.66 (0.41–6.77)	2.92 (0.57–15.12)	0.78 (0.17–3.55)
Maternal nativity
Foreign born (versus US born)	0.77 (0.66–0.89)	0.82 (0.72–0.94)	0.67 (0.56–0.80)	0.91 (0.81–1.03)
Maternal education
<12 years	0.94 (0.79–1.12)	0.93 (0.80–1.09)	1.17 (0.96–1.43)	1.11 (0.97–1.27)
12 years	Reference	Reference	Reference	Reference
>12 years	0.87 (0.73–1.03)	1.22 (1.06–1.42)	1.12 (0.92–1.37)	0.90 (0.78–1.03)
Borough of mother’s usual residence at birth
Manhattan	Reference	Reference	Reference	Reference
Bronx	1.04 (0.83–1.32)	0.81 (0.66–0.99)	1.03 (0.80–1.33)	0.92 (0.77–1.10)
Brooklyn	1.10 (0.89–1.36)	0.94 (0.78–1.12)	0.76 (0.60–0.97)	0.96 (0.81–1.14)
Queens	1.29 (1.02–1.63)	0.88 (0.72–1.07)	1.01 (0.77–1.31)	0.96 (0.80–1.15)
Staten Island	0.79 (0.56–1.11)	0.71 (0.54–0.94)	1.27 (0.89–1.81)	1.08 (0.83–1.40)
Maternal age
<25 years	0.87 (0.73–1.04)	0.98 (0.85–1.14)	0.81 (0.66–0.98)	1.11 (0.97–1.27)
25–34 years	Reference	Reference	Reference	Reference
≥35 years	1.32 (1.09–1.59)	0.99 (0.84–1.18)	1.23 (0.98–1.53)	1.15 (0.98–1.34)
Previous children
Nulliparous (versus multiparous)	1.15 (0.99–1.33)	1.10 (0.97–1.25)	1.50 (1.27–1.77)	1.16 (1.04–1.31)
Maternal obesity
Yes (versus no)	1.17 (0.92–1.49)	1.08 (0.87–1.34)	0.87 (0.65–1.16)	1.28 (1.06–1.56)
Child sex
Male (versus female)	1.20 (1.05–1.38)	1.05 (0.94–1.18)	2.70 (2.25–3.24)	1.16 (1.04–1.29)
Gestational age
≤31 weeks	2.30 (1.69–3.15)	2.00 (1.49–2.67)	0.75 (0.48–1.18)	0.36 (0.26–0.51)
32–33 weeks	1.38 (0.91–2.09)	1.54 (1.07–2.22)	0.85 (0.49–1.48)	0.72 (0.50–1.05)
34–36 weeks	1.32 (1.04–1.68)	1.22 (0.98–1.50)	1.05 (0.79–1.40)	0.83 (0.68–1.01)
37–38 weeks	1.17 (0.99–1.39)	1.10 (0.95–1.28)	1.25 (1.03–1.50)	0.99 (0.87–1.14)
39–40 weeks	Reference	Reference	Reference	Reference
≥41 weeks	1.10 (0.87–1.41)	1.12 (0.92–1.37)	1.21 (0.94–1.57)	1.12 (0.93–1.34)
Small for gestational age
Yes (versus no)	1.56 (1.32–1.86)	1.38 (1.18–1.61)	0.92 (0.75–1.14)	0.96 (0.83–1.12)
Child required NICU
Yes (versus no)	1.15 (0.93–1.43)	0.92 (0.76–1.12)	0.65 (0.65–1.10)	0.87 (0.73–1.05)
Child’s 5-minute Apgar score < 7
Yes (versus no)	1.65 (1.11–2.47)	1.28 (0.86–1.89)	0.85 (0.45–1.61)	0.74 (0.47–1.17)
Child had any congenital anomaly
Yes (versus no)	3.14 (1.94–5.06)	1.52 (0.93–2.48)	1.28 (0.65–2.52)	0.75 (0.45–1.26)

Table shows the results from multinomial regression modelling odds ratios for patterns 1, 2, 4, and 5 versus pattern 3. All variables in Tables 1–3 were included in the model except for maternal origin if foreign born, maternal hypertension and diabetes, and birth weight because they overlapped with other variables in the model. In this table, only variables with a Wald P value < 0.05 are shown. Statistically significant results are bolded

Children with multiple therapies across EI/ECSE/SE (versus pattern 3) were significantly more likely to have a Black (adjusted odds ratio [aOR] 1.27) or Latina (aOR 1.25) mother, have a mother living in Queens (aOR 1.29), have a mother ≥35 years (aOR 1.32), be male (aOR 1.20), have gestational age ≤31 weeks (aOR 2.30) or 34–36 weeks (aOR 1.32), be small for gestational age (aOR 1.56), have an Apgar < 7 (aOR 1.65) and have any congenital anomaly (aOR 3.14). They were significantly less likely to have a foreign-born mother (aOR 0.77).

Children with EI without transition to DOE schools or services (versus pattern 3) were significantly more likely to have a mother with > 12 years of education (aOR 1.22), have gestational age ≤31 weeks (aOR 2.00) or 32–33 weeks (aOR 1.54), and be small for gestational age (aOR 1.38). They were significantly less likely to have a foreign-born mother (aOR 0.82) or to have a mother living in the Bronx (aOR 0.81) or Staten Island (aOR 0.71).

Children with older entry to EI and both ST and OT throughout ECSE/SE (versus pattern 3) were significantly more likely to have a Black (aOR 2.05) or Latina (aOR 1.49) mother, be a first child (aOR 1.50), be male (aOR 2.70), and have a gestational age 37–38 weeks (aOR 1.25). They were significantly less likely to have a foreign-born mother (aOR 0.67), a mother living in Brooklyn (aOR 0.76), and a mother < 25 years (aOR 0.81).

Children with limited EI use and mostly ST in ECSE/SE (versus pattern 3) were significantly more likely to have a Black (aOR 1.41), Latina (aOR 1.43), or Asian/Pacific Islander (PI) (aOR 1.38) mother, be a first child (aOR 1.16), have a mother with obesity (aOR 1.28), and be male (aOR 1.16). They were significantly less likely to have a gestational age ≤31 weeks (aOR 0.36).

Discussion

Main findings of this study

In this cohort of children born to NYC resident mothers, we found that birth characteristics varied substantially by EI and SE service use patterns. Children in patterns 1 and 2 had the most adverse birth outcome profiles (e.g. premature and small for gestational age), which likely prompted their early referral and use of EI. For pediatricians who are often the first referral source to EI, it is particularly relevant that children in pattern 4, with older entry to EI and both ST and OT throughout ECSE/SE, had fewer pregnancy risk factors and better birth outcomes than the overall population of children who used IDEA services. Children in pattern 4 had a service use profile consistent with children with ASD—later entry into services, likely indicative of later diagnosis of ASD, and use of ST and OT.

The American Academy of Pediatrics (AAP) recommends developmental screening for children at 9, 18, and 24 or 30 months and developmental surveillance at all other well visits²⁵; however, many children do not receive adequate screening and/or surveillance.^26,27 Our findings highlight the importance of both parents and healthcare professionals in looking for early signs of developmental delays and disabilities and evaluating children for EI services, even among children with less adverse pregnancy and birth outcomes.

What is already known about this topic

The birth characteristics of children with different patterns of service use across the entire spectrum of EI, ECSE and SE have not been previously described. However, some birth characteristics associated with referral and/or entry into EI or SE and configuration of services have been reported:

Children of mothers of color have suboptimal provider screening and surveillance for developmental delays and lower EI referral rates than children of White mothers.^8,28–31 Recent qualitative research among Latina and Black mothers suggests that additional factors need to be addressed even when a developmental screen is completed, including maternal beliefs about developmental delays and EI and the influence of social networks and social/financial stressors.³² Race also appears to be relevant to the configuration of services, as does family income.¹²

Because EI and SE provide services for children with developmental delays and disabilities, birth characteristics of children with developmental delays and disabilities in general can also predict service use to some extent. For example, gestational age is robustly associated with neurodevelopment and long-term medical and social outcomes³³ and predicts timing of EI referral.³⁴ However, service use is influenced by multiple systematic barriers, which is why service use, even for children born extremely preterm, can vary.³⁵

What this study adds

This study provides insight into the birth characteristics of children with different patterns of EI and SE service use, with specific insights regarding maternal race, gestational age, parity, and maternal obesity.

Maternal race was especially relevant for children in patterns 4 and 5, who were more likely to have a Black or Latina mother compared with children in pattern 3, consistent with racial disparities observed in other studies.^8,28–31 Although children in pattern 5 (versus pattern 3) were significantly more likely to have an Asian/PI mother, overall, children of Asian/PI mothers less often used EI and SE, consistent with previous research.¹⁰ However, this appears to be changing in NYC based on the most recent data.³⁶ To better understand the reasons behind these racial disparities, further qualitative research would be beneficial to understand the barriers that families face to EI enrollment and service use.

For gestational age, we observed a dichotomy in our study; in our multinomial regression model, prematurity was a strong predictor for early entry to EI (i.e. patterns 1 and 2), which was expected as these children have the most adverse clinical profiles, but it was not a positive predictor of older entry to EI or limited use of EI, compared with children in pattern 3. In fact, a gestational age < 31 weeks was a negative predictor of limited EI use and mostly ST in ECSE/SE. These findings highlight that although early gestational age might be useful to identify children who will be early and intensive users of EI, it is less useful in identifying children who enter EI later or have limited EI use. Another finding of interest was that early term birth (i.e. 37–38 weeks) was a positive predictor of older entry to EI and both ST and OT throughout ECSE/SE compared with pattern 3, adding to the growing body of evidence that early term infants, including planned early term births, are at greater developmental risk compared with infants born at term.^37,38

Children in patterns 4 and 5 were significantly more likely to be a first-born child compared with children in pattern 3. This highlights the importance of developmental screening and surveillance for the children of first-time parents who do not have another child as a reference for normal development. The importance of regular well visits should be emphasized, especially to first-time parents, and education should be provided about developmental delays and disabilities in a way that is accessible and understandable for all parents and families.

Maternal obesity was only predictive of pattern 5, with children in this pattern more likely to have a mother with obesity compared with children in pattern 3. Although we did not have data on specific diagnoses, we hypothesize that children in pattern 5 may have required ST because of auditory processing problems, which were identified when these children started taking on academic tasks in pre-kindergarten and kindergarten. Conversely, children who received early EI, such as children in pattern 3, may have had problems with language acquisition. Previous work has found an association between maternal obesity and child developmental delays, including a study among NYC children that found an association with moderate-severe cognitive and physical delays.²² Other work also documents an association between maternal obesity and specific delays, especially cognitive delays.^39,40 A biological mechanism for the association between maternal obesity and developmental delays remains to be established, but there are some promising leads in epigenetics.⁴¹

Limitations of this study

The study population and timeliness of the data are perhaps the biggest limitations of our study. Although our study included records for all children born in NYC, it was limited to children who remained in NYC after birth, used the public-school system, and remained alive through third grade. Therefore, our findings cannot be generalized to all NYC children, let alone children outside of NYC. In the US, EI and SE are not administered at a federal level, so at the state level (or at the city level in the case of NYC), there are differences in eligibility criteria and procedures,⁴² which likely reflect differences in service use. However, federal law sets the minimum requirements by which individual jurisdictions must abide, so broadly speaking these differences in eligibility and procedures are not extreme. This cohort of children was born in 1998 and followed through the 2006–2007 school year. Timeliness of service use data is important because of the changing epidemiology of diagnosed childhood disabilities, such as ASD, changing attitudes regarding childhood disabilities, and changes in EI/ECSE/SE programs themselves. Therefore, it will be important to repeat this analysis as more recent data become available.

Conclusions

The public health implications of our findings are twofold. First, clinicians involved in developmental screening and surveillance can use these findings to better identify children who may be at risk of suboptimal service use. Specifically, they can be especially vigilant for children of mothers of color and children who are first born and children who may have less adverse birth characteristic profiles (e.g. no congenital anomalies, no NICU use, etc.) as these children may indeed require services during childhood, albeit less intensive than children with the most adverse birth characteristic profiles. Second, regarding policy, this information may be helpful to optimize existing EI services. Based on the risk profiles identified, children who are at risk for suboptimal service use can be targeted for earlier identification (e.g. via clinician education and outreach) and more intensive service use (e.g. via outreach to EI service providers).

Funding source

This study was supported by the New York City Department of Health and Mental Hygiene. Dr. Stingone was supported by a grant from the NIH/NIEHS (ES027022).

Contributors

Dr. Romo carried out the analysis, drafted the initial manuscript, and reviewed and revised the manuscript; Dr. McVeigh conceptualized and designed the study, contributed to the interpretation of data and critically reviewed the manuscript for important intellectual content; Ms Jordan, Dr. Stingone, Ms Chan and Dr. Askew contributed to the interpretation of data and critically reviewed the manuscript for important intellectual content; all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

Human participant protection

This analysis was approved by the New York City Department of Health and Mental Hygiene Institutional Review Board.

Appendix 1. Definitions of maternal medical risk factors, labor or delivery complications, abnormal conditions of the newborn and congenital anomalies

Maternal medical risk factors: anemia, cardiac disease, acute or chronic lung disease, diabetes (gestational, chronic), genital herpes, other STD, hydramnios/oligohydramnios, hemoglobinopathy, hepatitis, hypertension (chronic, pregnancy-associated), pre-eclampsia, eclampsia, incompetent cervix, previous infant 4000+ g, previous preterm or small for gestational age infant, renal disease, RH sensitization, uterine bleeding (trimesters 1, 2 and 3), genetic disease, seizure disorders, previous spontaneous abortion, thrombophlebitis, thyroid condition, in vitro fertilization, treatment for infertility, rubella, tuberculosis, viral disease, and other.

Labor or delivery complications: anesthetic complications, abruption placenta, placenta previa, cord prolapse, conditions of cord, fetal distress, cephalopelvic disproportion, chorioamnionitis, meconium staining, premature rupture (>12 hours), seizures during labor, precipitous labor (<3 hours), prolonged labor (>20 hours), failure to progress, breech/malpresentation, febrile (>100 F), coagulation defects hemorrhage postpartum, lacerations (cervical or vaginal), marginal sinus rupture, uterine atony, retained placenta, premature rupture (1–12 hours), other.

Abnormal condition of the newborn: anemia, birth injury, fetal alcohol syndrome, hyaline membrane disease/RDS, meconium aspiration syndrome, assisted ventilation (intubation, other), seizures, congenital infections, drug withdrawal syndrome, metabolic disorders, RDS, and other.

Possible neonatal hypoxic condition: neonatal respiratory distress syndrome, hyaline membrane disease, meconium aspiration syndrome, anemia, assisted ventilation (intubation), assisted ventilation (other), fetal distress syndrome.

Congenital anomaly: anencephalus, neural tube defects, hydrocephalus, microcephalus, other CNS anomalies, heart malformations, other circulatory/respiratory, rectal atresia/stenosis, tracheo-esophageal fistula–atresia, omphalocele–gastroschisis, other GI anomalies, malformed genitalia, renal anomalies, other urogenital anomalies, clef lip/palate, polydactyly/syndactyly/adactyly, club foot, diaphragmatic hernia, other musculoskeletal/integumental, trisomy 21, other chromosomal, encephalocele, single umbilical artery, hydronephrosis, limb reduction, congenital rubella syndrome, and other.

Matthew L. Romo, Consultant

Katharine H. McVeigh, Director of Research

Phoebe Jordan, Graduate Student

Jeanette A. Stingone, Assistant Professor

Pui Ying Chan, City Research Scientist

George L. Askew, Former Deputy Commissioner

Conflict of interest

All authors have indicated they have no potential conflicts of interest to disclose.