Prevalence of cerebral palsy, intellectual disability, hearing loss, and blindness, National Health Interview Survey, 2009–2016

Abstract

Background:

Developmental disabilities are present in a significant proportion of US children. Surveillance of developmental disabilities is crucial for monitoring population trends, guiding research into risk factors, and informing resource allocation.

Objective/Hypothesis:

We examined overall prevalence, prevalence by demographic characteristics, and trends over time for cerebral palsy (CP), intellectual disability (ID), moderate to severe hearing loss (MSHL), and blindness.

Methods:

Data from the 2009–2016 National Health Interview Survey (NHIS) were analyzed for children 3–17 years of age. Question wording was consistent over time except for ID, which changed in 2011 to replace the term “mental retardation.” Demographic differences and linear trends (over three time periods) were assessed by Chi-square tests and Wald-F tests.

Results:

Prevalence estimates per 1000 children ages 3–17 years for CP, ID, MSHL, and blindness were 3.2 (95% CI: 2.7, 3.7), 11.1 (95% CI: 10.2, 12.1), 6.4 (95% CI: 5.6, 7.2), and 1.6 (95% CI: l.3, 2.0), respectively. Disability prevalence was higher for children with low birth weight and from families of lower parental education, income ≤200% of federal poverty level, and public insurance. Older children had higher ID prevalence; boys had significantly higher CP and ID prevalences. Only ID demonstrated a significantly increased trend over time (p = 0.0002).

Conclusions:

We provide nationally representative prevalence estimates for four developmental disabilities; recent estimates are comparable to those from records-based studies. Prevalences were stable except for ID, which increased after 2010, coincident with the questionnaire change. A substantial number of US children continue to have these disabilities and service needs.

Introduction

According to the most recent national data, 17% of children in the United States have a developmental disability,¹ a collection of chronic conditions originating in childhood manifested as physical, psychological, cognitive, and speech impairments.² Developmental disabilities typically persist throughout a child’s life and therefore require significant resources and planning for support services and assistive technology over the life span for optimal health, education, and functioning. As compared to children without disabilities, children with disabilities require more physician visits and hospitalizations, and their parents report more school days lost and poorer educational outcomes.³ Children with at least one developmental disability have higher rates of health care utilization and more often need help with personal care, require special equipment and home health services, and have increased prescription medication use.⁴ Timely surveillance of developmental disabilities is crucial for examining risk factors, developing effective programs and policies, and providing early intervention to maximize health, well-being, and quality of life.² Additionally, current prevalence estimates and regular reporting are essential for estimating the costs of programs and policies, and facilitating prioritization and modification in resource allocation to ensure that proper supports and services are available as needs change.⁵

Cerebral palsy (CP), intellectual disability (ID), moderate to severe hearing loss (MSHL), and blindness are four disabilities that usually affect functioning of individuals throughout their lifetime. CP describes a group of permanent, non-progressive disorders of the development of movement and posture. It is the most common physical disability of childhood^6,7 and is often accompanied by disturbances in sensation, perception, cognition, communication, and behavior as well as epilepsy and secondary musculoskeletal problems.⁸ ID originates before 18 years of age and is characterized by significant limitations in intellectual functioning and in adaptive behavior, which includes conceptual, social, and practical skills used every day.⁹ Moderate hearing loss is the ability to hear sounds at 41 dB or greater; severe hearing loss is the ability to hear sounds at 71 dB or greater.¹⁰ Blindness (as opposed to vision impairment or low vision) is the inability to see anything with either eye.¹¹ While hearing loss and blindness that occur as isolated disabilities may be associated with greater functional level,³ like CP and ID, they often occur in conjunction with each other and other developmental disabilities, which substantially compounds the functional impact on the child.¹² For individuals born in 2000, average lifetime costs were estimated in 2003 dollars at $921,000 for persons with cerebral palsy, $1,014,000 for persons with mental retardation, $383,000 for persons with hearing loss, and $601,000 for persons with vision impairment.¹³ Regular surveillance of these four developmental disabilities is essential for understanding their characteristics and assisting children and their families in receiving the interventions and supports needed to maximize developmental, academic, and transitional outcomes.^14,15

Nationally representative estimates were last reported for CP and ID in 2016 using 2011–2013 data,¹⁶ and the most recent reports for hearing loss and blindness were published in 2011 from 1997 to 2008 data.¹⁷ The purpose of this study was to provide updated prevalence estimates of CP, ID, MSHL, and blindness in children ages 3–17 years using data from a national health survey. We examined overall prevalence, prevalence by demographic characteristics, and trends over time.

Methods

Data from the 2009–2016 National Health Interview Survey (NHIS) were used for the analysis. NHIS is a nationally representative survey of the civilian, non-institutionalized US population. One sample child is randomly selected from each family with children. Through in-person, computer-assisted interviews with a knowledgeable parent or guardian, information about the child (sample child data file) is collected and supplemented with basic health and demographic information from the household (family and person data files). The current analysis only included children ages 3–17 years because many developmental disabilities are not yet diagnosed in younger children (81,133 children were 3–17 years of age out of the 98,196 children in the sample child file). During 2009–2016, the sample child final response rate ranged from 62% (2016) to 75% (2011). Further details on NHIS surveys and sampling descriptions are available at: https://www.cdc.gov/nchs/nhis/data-questionnaires-documentation.htm.

The four outcomes assessed included CP, ID, MSHL, and blindness. A child was considered to have one of these disabilities if there was an affirmative response to the following questions asked of parents and/or guardians: “Has a doctor or other health professional ever told you that [Sample Child] had cerebral palsy?” (CP); “Has a doctor or health professional ever told you that [Sample Child] had mental retardation?” (2009–2010) or “Has a doctor or health professional ever told you that [Sample Child] had an intellectual disability, also known as mental retardation?” (2011–2016) (ID); “Which statement best describes [Sample Child’s] hearing without a hearing aid: excellent, good, a little trouble hearing, moderate trouble, a lot of trouble, or is [Sample Child] deaf?’ (MSHL); and “Is [Sample Child] blind or unable to see at all?” (blindness). MSHL was dichotomized so that an affirmative response included the last three categories: moderate trouble hearing, a lot of trouble hearing, or deafness. Except for ID, all questions were asked in the same manner over the eight survey years.

Data were weighted to produce national estimates of the non-institutionalized population of children ages 3–17 years. Prevalence of CP, ID, MSHL, and blindness were examined overall and by demographic characteristics including the child’s age (3–10 years and 11–17 years), sex, race/ethnicity (white, non-Hispanic; black, non-Hispanic; Hispanic; and other), highest maternal or paternal educational attainment (less than high school, high school or some college, and college degree or higher), poverty level defined as family income relative to the poverty threshold (less than as compared to greater th.an or equal to 200%), health insurance coverage (private only, public including state and governmental programs, and uninsured), and birthweight (normal (≥2500 g), low (<2500 g, LBW), and very low (<1500 g, VLBW)). Parents were able to report birth weight in pounds and ounces, or grams; we used the variable provided with weights converted to grams. Demographic variables were categorized similarly to Boyle et al. for comparison where possible.¹⁷ However, insurance coverage was collapsed into three mutually exclusive categories similarly to Radcliff et al. for ease in interpreting the results¹⁸ whereas birthweight categories were not mutually exclusive but grouped according to standard convention.¹⁹ We compared LBW (<2500 g) and VLBW (<1500 g) to normal birthweight (≥2500 g). Prevalence of any disability, having at least one of the four disabilities, was also estimated overall and by demographic characteristics. We did not analyze the co-occurrence of the four disabilities as we did not have the statistical power to do so.

A proportion of observations had missing demographic data: parental education 0.2%, insurance coverage 0.4%, and birthweight 10.4%. These observations were excluded from the stratified analyses. Multiply imputed income files provided by the National Center for Health Statistics were used to avoid potentially biased estimates due to missing poverty observations. Chi-square tests were used to examine prevalence differences between demographic subgroups. Wald-F tests were used to assess linear trends over three constructed time periods (2009–2010, 2011–2013, and 2014–2016) and the differences in the trends between the periods. Only the years within each specific time period were used in the analyses and were weighted accordingly. Estimates with a relative standard error greater than 30% and less than or equal to 50% were indicated as such in Tables 1, 2 and 4 and should be interpreted with caution as they do not meet standards of reliability or precision; estimates with a relative standard error greater than 50% were also indicated as such in Tables 1 and 4 and were suppressed.²⁰ P values less than 0.05 were considered statistically significant. SUDAAN, release 11.0 (Research Triangle Institute International, Research Triangle Park, NC) was used to account for the complex survey design.

Table 1.

Weighted prevalence of cerebral palsy^a in children 3–17 years of age (N = 81,088) by selected demographic characteristics, NHIS, 2009–2016.

	Unweighted Noncases	Unweighted Cases	Prevalenceb [95% CI]	P valuec
Overall	80,843	245	3.2 [2.7, 3.7]
Age (Years)
3–10 years	40,961	124	2.8 [2.2, 3.6]	0.16
11–17 years	39,882	121	3.6 [2.8, 4.5]
Sex
Girls	39,290	88	2.2 [1.7, 2.9]	0.0004
Boys	41,553	157	4.1 [3.3, 5.0]
Race/Ethnicity
White, non-Hispanic	36,917	123	3.2 [2.5, 4.0]	0.23
Black, non-Hispanic	11,966	46	4.3 [2.8, 6.3]
Hispanic	23,201	56	2.4 [1.8, 3.4]
Otherd	8759	20	3.3 [1.6, 6.7]
Highest Parental Educatione
Less than HS	8655	24	2.7 [1.7, 4.5]f	0.82
HS or Some College	43,703	135	3.2 [2.6, 4.0]
College Graduate or Higher	28,362	86	3.2 [2.4, 4.3]
Poverty Levelg
<200%	35,924	132	3.9 [3.1, 4.9]	0.03
≥200%	44,919	113	2.6 [2.0, 3.3]
Health Insurance Coverage
Privateh	44191	102	2.5 [1.9, 3.2]	0.0001
Publici	30125	136	4.6 [3.7, 5.8]
Uninsuredj	6181	7	k
Birthweight
Normal (≥2500 g)	65,596	114	1.8 [1.4, 2.3]
Low (<2500 g)	6879	113	16.8 [13.3, 21.0]	<0.0001l
Very Low (<1500 g)	1455	67	48.1 [36.0, 64.1]	<0.0001m

CI = confidence interval; HS = high school; g = grams.

^aCP assessed by parent answering “Has a doctor or other health professional ever told you that [Sample Child] had cerebral palsy?”

^bWeighted prevalence per 1000.

^cChisq test between cerebral palsy and each demographic variable.

^dOther race includes American Indian, Alaskan Native; Asian; and multiple races.

^eIndicates highest maternal or paternal educational attainment.

^fEstimate with a relative standard error greater than 30% and less than or equal to 50% and should be used with caution as it does not meet standards of reliability or precision.

^gFamily income level relative to the 200% poverty threshold.

^hIncludes persons who had any comprehensive private insurance plan (including health maintenance organizations, preferred provider organizations, and exchange-based coverage) and those with Tricare, CHAMPS-VA or other military insurance plans excluding VA coverage.

ⁱIncludes persons who do not have any type of private coverage, but who have Medicaid, other state-sponsored health plans (including CHIP), Medicare, VA coverage, or other governmental plans.

^jIncludes persons who have not indicated that they are covered under private health insurance, Medicare, Medicaid, CHIP, a state-sponsored health plan, other government programs, or military coverage. This category also includes person who are covered by Indian Health Service only or who only have a plan that pays for one type of service such as accidents or dental care.

^kEstimate with a relative standard error > 50% and therefore not shown.

^lChisq test between normal and low birthweight which includes very low birthweight; birthweight levels are not mutually exclusive.

^mChisq test between normal and very low birthweighr; birthweight levels are not mutually exclusive.

Table 2.

Weighted prevalence of intellectual disability^a in children 3–17 years (N = 81,084) or age by selected demographic characteristics, national health interview survey, 2009–2016.

	Unweighted Noncases	Unweighted Cases	Prevalenceb [95% CI]	P valuec
Overall	80,153	931	11.1 [10.2, 12.1]
Age (Years)
3–10 years	40,747	337	8.1 [7.1, 9.2]	<0.0001
11–17 years	39,406	594	14.0 [13.1, 16.3]
Sex
Girls	39,016	358	7.8 [6.9, 9.0]	<0.0001
Boys	41,137	573	14.2 [12.8, 15.9]
Race/Ethnicity
White, non-Hispanic	36,606	434	10.6 [9.5, 11.9]	0.06
Black, non-Hispanic	11,842	168	14.2 [11.8, 17.1]
Hispanic	22,999	258	11.3 [9.6, 13.2]
Otherd	8706	71	8.8 [6.1, 12.8]
Highest Parental Educatione
Less than HS	8538	140	16.3 [13.3, 19.9]	<0.0001
HS or Some College	43,303	527	11.9 [10.6, 13.3]
College Graduate or Higher	28,188	264	8.7 [7.4, 10.2]
Poverty Levelf
<200%	35,512	542	15.2 [13.5, 17.1]	<0.0001
≥200%	44,641	389	8.0 [7.1, 9.0]
Health Insurance Coverage
Privateg	43946	340	7.3 [6.4, 8.2]	<0.0001
Publich	29706	554	18.2 [16.2, 20.3]
Uninsuredi	6160	31	5.8 [3.2, 10.3]j
Birthweight
Normal (≥2500 g)	65,071	636	9.4 [8.5, 10.5.0]
Low (<2500 g)	6826	167	23.3 [19.4, 28]	<0.0001k
Very Low (<1500 g)	1460	64	38.7 [29.1, 51.3]	<0.0001l

CI = confidence interval; HS = high school; g = grams.

^aIntellectual Disability assessed by parent answering “Has a doctor or health professional ever told you that [Sample Child] had mental retardation?” (2009–2010) or “Has a doctor or health professional ever told you that [Sample Child] had an ID, also known as mental retardation?” (2011–2016).

^bWeighted prevalence per 1000.

^cChisq test between intellectual disability and each demographic variable.

^dOther race includes American Indian, Alaskan Native; Asian; and multiple races.

^eIndicates highest maternal or paternal educational attainment.

^fFamily income level relative to the 200% poverty threshold.

^gIncludes persons who had any comprehensive private insurance plan (including health maintenance organizations, preferred provider organizations, and exchange-based coverage) and those with Tricare, CHAMPS-VA or other military insurance plans excluding VA coverage.

^hIncludes persons who do not have any type of private coverage, but who have Medicaid, other state-sponsored health plans (including CHIP), Medicare, VA coverage, or other governmental plans.

ⁱIncludes persons who have not indicated that they are covered under private health insurance, Medicare, Medicaid, CHIP, a state-sponsored health plan, other government programs, or military coverage. This category also includes person who are covered by Indian Health Service only or who only have a plan that pays for one type of service such as accidents or dental care.

^jEstimate with a relative standard error greater than 30% and less than or equal to 50% and should be used with caution as it does not meet standards or reliability or precision.

^kChisq test between normal and low birthweight which includes very low birthweight, birthweight levels are not mutually exclusive.

^lChisq test between normal and very low birthweight, birthweight levels are not mutually exclusive.

Table 4.

Weighted prevalence of blindness^a in Children 3–17 years (N = 81,053) or age by selected demographic characteristics, national health interview survey. 2009–2016.

	Unweighted Noncases	Unweighted Cases	Prevalenceb [95% CI]	P valuec
Overall	80,927	126	1.6 [1.3, 2.0]
Age (Years)
3–10 years	41,000	56	1.3 [0.9, 1.9]	0.16
11–17 years	39,927	70	1.9 [1.4, 2.6]
Sex
Girls	39,289	65	1.7 [1.2, 2.4]	0.61
Boys	41,638	61	1.5 [1.1, 2.0]
Race/Ethnicity
White, non-Hispanic	36,975	49	1.4 [0.9, 2.0]	0.38
Black, non-Hispanic	11,989	17	1.4 [0.8, 2.6]
Hispanic	23,197	50	2.2 [1.5, 3.2]
Otherd	8766	10	e
Highest Parental Educationf
Less than HS	8655	20	2.7 [1.6, 4.5]g	0.04
HS or Some College	43,730	74	1.7 [1.3, 2.4]
College Graduate or Higher	28,418	32	1.1 [0.7, 1.6]
Poverty Levelh
<200%	35,946	83	2.4 [1.8, 3.3]	0.0002
≥200%	44,981	43	0.9 [0.6, 1.2]
Health Insurance Coverage
Privatei	44253	43	0.9 [0.6, 1.3]	0.0006
Publicj	30153	73	2.7 [1.9, 3.8]
Uninsuredk	6174	10	e
Birthweight
Normal (≥2500 g)	65,580	88	1.2 [1.0, 1.6]
Low (<2500 g)	6962	29	4.9 [2,8, 8.7]	0.01l
Very Low (<1500 g)	1501	22	19.2 [10.2, 35.7]g	0.004m

CI = confidence interval; HS = high school; g = grams.

^aBlindness assessed by parent answering “Is [Sample Child] blind or unable to see at all?.

^bWeighted prevalence per 1000.

^cChisq test between blindness and each demographic variable.

^dOther race includes American Indian, Alaskan Native; Asian; and multiple races.

^eEstimate with a relative standard error > 50% and therefore are not shown.

^fIndicates highest maternal or paternal educational attainment.

^gEstimate with a relative standard error greater than 30% and less than or equal to 50% and should be used with caution as it does not meet standards or reliability or precision.

^hFamily income level relative to the 200% poverty threshold.

ⁱIncludes persons who had any comprehensive private insurance plan (including health maintenance organizations, preferred provider organizations, and exchange-based coverage) and those with Tricare, CHAMPS-VA or other military insurance plans excluding VA coverage.

^jIncludes persons who do not have any type of private coverage, but who have Medicaid, other state-sponsored health plans (including CHIP), Medicare, VA coverage, or other governmental plans.

^kIncludes persons who have not indicated that they are covered under private health insurance, Medicare, Medicaid, CHIP, a state-sponsored health plan, other government programs, or military coverage. This category also includes person who are covered by Indian Health Service only or who only have a plan that pays for one type of service such as accidents or dental care.

^lChisq test between normal and low birthweight which includes very low birthweight, birthweight levels are not mutually exclusive.

^mChisq test between normal and very low birthweight, birthweight levels are not mutually exclusive.

Results

Among 81,133 children 3–17 years of age, 81,088 children had data on CP; 81,084 had data on ID; 81,090 had data on MSHL; and 81,053 had data on blindness. During 2009–2016, the overall prevalence estimates per 1000 children ages 3–17 years for CP, ID, MSHL, and blindness were 3.2 (95% CI: 2.7, 3.7), 11.1 (95% CI: 10.2, 12.1), 6.4 (95% CI: 5.6, 7.2), and 1.6 (95% CI: 1.3, 2.0), respectively (Tables 1–4). Older children (ages 11–17) had a higher prevalence of ID; otherwise, there were no statistically significant differences between younger and older children or by race/ethnicity group. Between boys and girls, boys had significantly higher CP and ID prevalences. Overall, children of parents with higher education had significantly lower prevalences of ID, MSHL, and blindness, and children from families in the lower poverty threshold category had significantly higher prevalences of all four outcomes.

Insurance coverage and birthweight demonstrated statistically significant associations with all disability outcomes (Tables 1–4). The majority of children were covered under private or public insurance programs. Specifically, at least 95% of children with public insurance who had any of the four developmental disabilities had Medicaid coverage. Prevalence was twice as high among children insured by public insurance as compared to those insured by private insurance, a statistically significant pattern for all outcomes. ID and MSHL prevalences were also significantly higher for children insured by public insurance as compared to those who were uninsured. VLBW was significantly associated with more than a threefold higher prevalence of all outcomes relative to normal birthweight; ID and MSHL were 3–4 times more common, while CP and blindness were 16–25 times more common.

Patterns of prevalence by demographic characteristics for children with at least one of the developmental disabilities were similar to those for each individual disability. Older children, boys, children of parents with lower education, children from families in the lower poverty threshold, children with public insurance, and those with very low birth weight had significantly higher prevalences of any disability (Table 5).

Table 5.

Weighted prevalence of any developmental disability^a in children 3–17 years of age (N = 80.969) by selected demographic characteristics, national health interview survey, 2009–2016.

	Unweighted Noncases	Unweighted Cases	Prevalenceb [95% CI]	P valuec
Overall	79,382	1587	19.6 [18.4, 21.0]
Age (Years)
3–10 years	40354	653	16.2 [14.7, 17.8]	<0.0001
11–17 years	39,028	934	23.6 [21.6, 25.8]
Sex
Girls	38,672	652	15.6 [14.1, 17.2]	<0.0001
Boys	40,710	935	23.5 [21.7, 25.6]
Race/Ethnicity
White, non-Hispanic	30,222	775	20.1 [18.4, 21.9]	.12
Black, non-Hispanic	11,721	264	22.2 [18.9, 26.1]
Hispanic	22,807	417	18.4 [16.3, 20.8]
Otherd	8632	131	16.3 [12.6, 21.0]
Highest Parental Educatione
Less than HS	8448	215	24.8 [21.1, 29.3]	<0.0001
HS or Some College	42,848	912	21.6 [19.9, 23.5]
College Graduate or Higher	27,964	460	15.6 [14.0, 17.5]
Poverty Levelf
<200%	35,065	914	26.2 [23.9, 28.7]	<0.0001
≥200%	44,317	673	14.5 [13.2, 15.8]
Health Insurance Coverage
Privateg	43,631	626	14.1 [12.8, 15.5]	<0.0001
Publich	29,308	880	29.7 [27.1, 32.5]
Uninsuredi	6104	74	13.2 [9.2, 18.9]
Birthweight
Normal (≥2500 g)	64,520	1105	16.9 [15.6, 18.3]
low (<2500 g)	6682	304	45.5 [39.6, 52.3]	<0.0001j
Very Low (<1500 g)	1387	134	93.9 [75.5, 116.3]	<0.000lk

CI = confidence interval; HS = high school; g = grams.

^aAny disability includes cerebral palsy, intellectual disability, moderate to severe hearing loss, or blindness.

^bWeighted prevalence per 1000.

^cChisq test between disability and each demographic variable.

^dOther race includes American Indian, Alaskan Native; Asian; and multiple races.

^eIndicates highest maternal or paternal educational attainment.

^fFamily income level relative to the 200% poverty threshold.

^gIncludes persons who had any comprehensive private insurance plan (including health maintenance organizations, preferred provider organizations, and exchange-based coverage) and those with Tricare, CHAMPS-VA or other military insurance plans excluding VA coverage.

^hIncludes persons who do not have any type of private coverage, but who have Medicaid, other state-sponsored health plans (including CHIP), Medicare, VA coverage, or other governmental plans.

ⁱIncludes persons who have not indicated that they are covered under private health insurance, Medicare, Medicaid, CHIP, a state-sponsored health plan, other government programs, or military coverage. This category also includes person who are covered by Indian Health Service only or who only have a plan that pays for one type of service such as accidents or dental care.

^jChisq test between normal and low birthweight which includes very low birthweight, birthweight levels are not mutually exclusive.

^kChisq test between normal and very low birthweight, birthweight levels are not mutually exclusive.

ID demonstrated a trend over the three constructed time periods with questionnaire wording 2009–2010: “Has a doctor or health professional ever told you that [Sample Child] had mental retardation?” 7.4 (6.1, 9.1); and 2011–2016: “Has a doctor or health professional ever told you that [Sample Child] had an intellectual disability, also known as mental retardation?” 2011–2013: 12.7 (11.3, 14.4); 2014–16: 11.9 (10.4, 13.7); p = 0.0003. Time trends were not seen in any of the other disabilities (Table 6).

Table 6.

Weighted prevalence of specific developmental disabilities in children 3–17 years of age by time period, national health interview survey, 2009–2016.

Disability	2009–2010	2011–2013	2014–2016	Overall Linear
	Prevalencea [95% CI]	Prevalencea [95% CI]	Prevalencea [95% CI]	Trendb
Cerebral Palsy	3.3 [2.4, 4.5]	3.0 [2.3, 3.8]	3.2 [2.4, 4.3]	.96
Intellectual Disability	7.4 [6.l, 9.1]c,d	12.7 [11.3, 14.4]e	11.9 [10.4, 13.7]	.0003
Hearing Loss	6.7 [5.3, 8.3]	6.7 [5.3, 8.4]	5.9 [4.8, 7.2]	.36
Blindness	1.4 [0.9, 2.1]	1.7 [1.1, 2.4]	1.6 [1.1, 2.4]	.67

CI = confidence interval.

^aWeighted prevalence per 1000.

^bTest of linear trend over the 3 time periods.

^cP < 0.0001, 2009–2010 versus 2011–2013.

^dP < 0.0001, 2009–2010 versus 2014–2016.

^eP = 0.50, 2011–2013 versus 2014–2016.

Discussion

This study provides the most up-to-date parent-reported prevalence estimates of CP, ID, MSHL, and blindness in children ages 3–17 years. Approximately 3.2 per 1000 children in the United States during 2009–2016 had a diagnosis of CP, 11.1 per 1000 children had ID, 6.4 per 1000 children had MSHL, and 1.6 per 1000 children had blindness. Older children were more likely to have ID, and boys were more likely to have a diagnosis of CP or ID. Children from families with incomes below 200% of the federal poverty level or with LBW had higher prevalences of all disabilities. A higher prevalence of children having ID, MSHL, and blindness were found in families in which neither the mother nor the father had a high school education. A higher proportion of children insured by Medicaid and other public or government-sponsored programs had a disability diagnosis as compared to those with private insurance coverage or those without coverage. The higher prevalence of disability under Medicaid coverage could be related to the twofold nature of eligibility criteria. Low-income families and qualified pregnant women and children are eligible for Medicaid.²¹ Additionally, children with a qualifying disability resulting in marked and severe functional limitation, which includes these four, may be eligible to receive Supplemental Security Income which also confers mandatory eligibility for Medicaid.²²

In general, these parent-reported findings were consistent with other national survey studies. CP prevalence (3.2 per 1000 children) was comparable to several studies (range between 2.6 and 3.9 per 1000 children)^{16,17,23–25} using national survey data, as the confidence interval included these estimates. Similar to the current study, Boyle et al. examined the prevalence of CP from 1997 to 2003 using data from NHIS.¹⁷ The authors reported a slightly higher mean CP prevalence of 3.9 per 1000, and patterns by demographic characteristics were somewhat different. They found a higher CP prevalence in girls and non-Hispanic white children as compared to the current study. Another study provided CP prevalence estimates in 2011–2013 NHIS data of 2.9 per 1000 (95% CI: 2.3, 3.7) and in 2011–2012 National Survey of Children’s Health (NSCH) data of 2.6 per 1000 (95% CI: 2.1, 3.2).¹⁶

Two U.S. records-based surveillance studies^23,25 also reported similar estimates for CP. Van Naarden Braun et al. reported a mean CP prevalence for 1991–2010 in metropolitan Atlanta of 3.5 per 1000,²³ and Durkin and colleagues calculated CP prevalence for the 2006, 2008, and 2010 surveillance years in four areas in the United States (central Alabama; metropolitan Atlanta, Georgia; metropolitan St. Louis, Missouri; and southeastern Wisconsin) at 3.5, 3.2, and 2.9 per 1000, respectively.²⁵ Like the current study, boys had a higher CP prevalence in these studies. While overall estimates were comparable to other studies, Durkin and colleagues²⁵ along with others in Europe (1980–2003)²⁶ and Australia (1970–2004)²⁷ found a decrease in CP over time. Like Boyle et al., we did not find this decrease. However, the current study had limited power to detect such a difference.

Like CP, ID has been similarly reported using national surveys and records-based surveillance. One study compared prevalence estimates between NHIS and NSCH during 2011–2013.¹⁶ Both of the surveys provided comparable prevalence estimates (12.1 and 12.2 per 1000). Using records-based surveillance, Van Naarden Braun et al. reported a comparable mean ID prevalence for 1991–2010 of 13.0 per 1000.²³ However, the current study’s prevalence estimates were higher than in a similar study that examined ID prevalence for the 12 years prior to the current study period.¹⁷ Boyle and colleagues reported a mean prevalence between 1997 and 2008 of almost half (7.1 per 1000) that of the current study.¹⁷ Of note, all surveys during the 12 year study period used the term ‘mental retardation’ as opposed to ‘intellectual disability’. In our current study, ‘mental retardation’ was only used in 2009 and 2010 before the terminology was updated to intellectual disability. It is likely that the terminology used led to the change in prevalence estimates, as prevalence estimates in 2009 and 2010 (7.8 and 7.1 per 1000, respectively) closely aligned with those reported by Boyle et al.¹⁷

The results for MSHL are least consistent with previous estimates, which may be explained in part by differences in hearing loss definitions. Two previous studies reported moderate-to-profound hearing loss estimates during the last 20 years.^17,23 One study examining NHIS data between 1997–2008¹⁷ reported a mean prevalence of 4.5 per 1000 (a 40% difference from 6.4 per 1000 children in the current study) and an overall decline in prevalence, yet demographic patterns were similar. Potential reasons for these differences are not clear but may be due to revised NHIS hearing loss categories in 2008. From 2008 to the present, MSHL has included children with moderate trouble hearing, a lot of trouble hearing, and those who were deaf. Prior to this time, the category only included children with a lot of trouble hearing and deafness—a more limited definition.

An even lower mean prevalence estimate, 1.4 per 1000 children from 1991 to 2010, was reported by a records-based surveillance system in metropolitan Atlanta that systematically reviewed clinical and educational records.²³ Unlike parent-reported data from NHIS, in the records-based study hearing loss was defined according to hearing test results and defined as bilateral measured loss of 40 dB or greater. This restrictive case definition may explain the much lower prevalence in these data as compared to the current and previous NHIS studies.

The blindness prevalence estimates were slightly lower than the range of other studies’ findings. NHIS estimates for the 12 years directly prior to the current study (1997–2008) ranged between 1.1 to 1.5¹⁷ per 1000 children; records-based estimates over 1991–2010 ranged between 0.7 and 1.0²³ per 1000 children. When compared to previous NHIS data, patterns within demographic characteristics were similar: children from families with lower education attainment and families with lower income levels had a higher proportion ofblindness.¹⁷ Overall findings were also similar to Van Naarden Braun et al. even though the methods and categorization of vision impairment differed between the studies.²³ Similar to MSHL, data from the study by Van Naarden Braun et al. were records-based as compared to parent-reported survey data gathered from NHIS. Additionally, Van Naarden Braun et al. used objectively defined test scores to define categories of blindness and visual impairment. Estimates included children who were either legally blind or who had low vision (visual acuity from 20/70 to 20/199 in the better-seeing eye that cannot be corrected or improved with regular eyeglasses).²³ Only 68% of the children in their study were legally blind. In the current study, children who were reported to be blind or unable to see at all were included, a more restricted categorization but subject to a range of possible parental interpretations as compared to the records-based study.

Up-to-date reporting of developmental disability prevalence is important for understanding overall and subgroup populations with disabilities to efficiently target and prioritize supports, interventions, and policies, and to monitor their effectiveness. The overall prevalences of CP, MSHL, and blindness were stable. ID was lower during 2009 and 2010, but similar for the rest of the time period. Except for CP, children of parents with a lower education level were more likely to have a disability; children of parents with a college degree were more likely to have CP than children of parents with lower education levels. The rationale for this finding is unknown although Boyle et al. demonstrated the same results examining the 12 years prior.¹⁷ Additionally, all outcomes were significantly associated with low birthweight. Due to changes in perinatal care it is important to monitor birthweight; however, caution is needed due to the limitations its assessment. While studies have shown a high correlation between parent-reported and records-based birthweight,^29,29 it may be more difficult to correctly classify children into LBW and VLBW categories based on parent report as compared to records-based birth weight.

These results are subject to at least four limitations. First, NHIS is administered through in-person interviews with the child’s parent or guardian and is therefore dependent on recall, comprehension, and interpretation of the questions asked. However, current findings were comparable to those of studies that utilized review of clinical and educational records, except for MSHL.^23–25 Second, even with combining eight years of data, the surveys had limited statistical power to precisely estimate prevalence for specific demographic subgroups, disability co-occurrence, and trends over time. ID was the only disability to demonstrate a statistically significant trend. However, this is likely due to the lower prevalence of ID during 2009 and 2010 when ‘mental retardation’ was the terminology used in the survey questionnaire. Parents may have been more likely to endorse ID in recent survey years due to the reduced stigma as compared to ‘mental retardation’ or due to misinterpretation of the term ‘intellectual disability’ to include other disabilities such as learning disabilities. Furthermore, recent prevalence estimates may be more accurate with the use of the term ID as they are similar to records-based estimates that use test-based scores. Third, with the exception of MSHL, survey questions simply asked whether the disability was present—not about the severity level. With this limited granularity, it is difficult to determine the full impact of the child’s disability. And although hearing loss allowed for several response levels, data collected were subjective. Parents’ subjective classifications might differ from clinical diagnoses based on objective decibel levels. Lastly, NHIS samples represent noninstitutionalized children and therefore a small percentage of children with more severe disabilities might not be accounted for.

While essential to public health, the surveillance of developmental disabilities through national survey administration is changing. NHIS is currently undergoing a significant redesign for the 2019 survey year to reduce respondent burden and align content with other federal health surveys.³⁰ With these changes, questions on CP will no longer be included in NHIS. Moving forward, NSCH will be the only survey that provides prevalence estimates for children with CP. Future estimates for CP may be consistent with previous estimates though as NHIS and NSCH have previously provided similar overall CP prevalence estimates.¹⁶

Conclusions

This analysis provides up-to-date, nationally representative prevalence estimates of CP, ID, MSHL, and blindness in children ages 3–17 years. Additionally, we were able to examine trends over time when questionnaire wording stayed the same. We found that overall prevalences of CP, MSHL, and blindness were stable during the last eight years. Prevalence of intellectual disability increased after 2010, coincident with the questionnaire wording change. Recent estimates for three of the four disabilities are comparable to those from records-based studies. The estimated prevalences of these disabilities, which have test-based or physical findings, have remained relatively constant over time as compared to prevalence estimates for those with behavioral characteristics such as autism spectrum disorder and attention-deficit/hyperactivity disorder, which have markedly increased.^16,17 A substantial number of US children continue to have these disabilities and service needs. Continued surveillance would provide a larger sample size to examine more accurately demographic characteristics, trends over time, co-occurrence of the disabilities, and information on severity for effective program and policy planning.

Table 3.

Weighted prevalence of moderate to severe hearing loss^a in children 3–17 years of age (N = 81,090) by selected demographic characteristics, national health interview survey, 2009–2016.

	Unweighted Noncases	Unweighted Cases	Prevalenceb [95% CI]	P valuec
Overall	80,593	497	6.4 [5.6, 7.2]
Age (Years)
3–10 years	40,853	229	6.0 [5.1, 7.2]	0.33
11–17 years	39,740	268	6.8 [5.8, 7.9]
Sex
Girls	39,145	225	5.7 [4.7, 6.8]	0.06
Boys	41,448	272	7.0 [6.0, 8.2]
Race/Ethnicity
White, non-Hispanic	36,778	262	7.0 [5.9, 8.3]	0.11
Black, non-Hispanic	11,948	69	5.8 [4.3, 8.0]
Hispanic	23,147	109	5.1 [4.1, 6.4]
Otherd	8720	57	6.8 [4.7, 9.8]
Highest Parental Educatione
Less than HS	8614	64	7.6 [5.5, 10.4]	0.004
HS or Some College	43,537	293	7.3 [6.2, 8.6]
College Graduate or Higher	28,319	140	4.8 [3.9, 5.9]
Poverty Levelf
<200%	35,755	296	8.8 [7.5, 10.3]	<0.0001
≥200%	44,838	201	4.5 [3.6, 5.5]
Health Insurance Coverage
Privateg	44092	209	4.7 [3.9, 5.7]	<0.0001
Publich	29996	258	9.2 [7.8, 10.9]
Uninsuredi	6159	29	5.1 [3.0, 8.9]
Birthweight
Normal (≥2500 g)	65,329	377	6.1 [53, 7.1]
Low (<2500 g)	6917	75	11.5 [8.4, 15.8]	0.005j
Very Low (<1500 g)	1497	27	19.8 [11.7, 33.1]	0.01k

CI = confidence interval; HS = high school; g = grams.

^aModerate to severe hearing loss assessed by parent answering “Which statement best describes [Sample Child’s] hearing without a hearing aid: excellent, good, a little trouble hearing, moderate trouble, a lot of trouble, or is [Sample Child] deaf?.

^bWeighted prevalence per 1000.

^cChisq test between hearing loss and each demographic variable.

^dOther race includes American Indian, Alaskan Native; Asian; and multiple races.

^eIndicates highest maternal or paternal educational attainment.

^fFamily income level relative to the 200% poverty threshold.

^gIncludes persons who had any comprehensive private insurance plan (including health maintenance organizations, preferred provider organizations, and exchange-based coverage) and those with Tricare, CHAMPS-VA or other military insurance plans excluding VA coverage.

^hIncludes persons who do not have any type of private coverage, but who have Medicaid, other state-sponsored health plans (including CHIP), Medicare, VA coverage, or other governmental plans.

ⁱIncludes persons who have not indicated that they are covered under private health insurance, Medicare, Medicaid. CHIP, a state-sponsored health plan, other government programs, or military coverage. This category also includes person who are covered by Indian Health Service only or who only have a plan that pays for one type of service such as accidents or dental care.

^jChisq test between normal and low birthweight which includes very low birthweight, birthweight levels are not mutually exclusive.

^kChisq test between normal and very low birthweight, birthweight levels are not mutually exclusive.