Overweight and Obesity Among Adults With Intellectual Disabilities Who Use Intellectual Disability/Developmental Disability Services in 20 U.S. States

Roger J. Stancliffe University of Sydney, Australia

K. Charlie Lakin and Sheryl Larson Research and Training Center on Community Living, University of Minnesota

Joshua Engler, Julie Bershadsky, and Sarah Taub Human Services Research Institute, Cambridge, MA

Jon Fortune Human Services Research Institute, Tualatin, OR

Renata Ticha Research and Training Center on Community Living, University of Minnesota

Abstract The authors compare the prevalence of obesity for National Core Indicators (NCI) survey participants with intellectual disability and the general U.S. adult population. In general, adults with intellectual disability did not differ from the general population in prevalence of obesity. For obesity and overweight combined, prevalence was lower for males with intellectual disability than for the general population but similar for women. There was higher prevalence of obesity among women with intellectual disability, individuals with Down syndrome, and people with milder intellectual disability. Obesity prevalence differed by living arrangement, with institutional residents having the lowest prevalence and people living in their own home the highest. When level of intellectual disability was taken into account, these differences were reduced, but some remained significant, especially for individuals with milder disability.

DOI: 10.1352/1944-7558-116.6.401

Overweight and obesity are associated with increased mortality and morbidity (Berrington de Gonzalez et al., 2010; Manson & Bassuk, 2003; Soverini et al., 2010). Obese individuals had a significantly higher mortality rate in a large sample of people with Down syndrome (Yang et al., 2002).

Among U.S. adults in the general population, the prevalence of obesity (body mass index [BMI]

$ 30.0) and overweight and obesity (BMI $ 25.0) in 2007–2008 was 33.8% and 68.0%, respectively (Flegal, Carroll, Ogden, & Curtin, 2010). There were important prevalence differences by gender, age group, and race–ethnic group. Obesity preva- lence among U.S. adults increased from 13%–15% in the 1960s and 1970s to 31% in 2000, but the rate of increase may now be leveling off (Flegal et al.).

VOLUME 116, NUMBER 6: 401–418 | NOVEMBER 2011 AJIDD

E American Association on Intellectual and Developmental Disabilities 401

Prevalence of obesity among adults with intellec- tual disabilities also increased between the mid- 1980s and 2000 (Rimmer & Yamaki, 2006; Yamaki, 2005).

Adults With Intellectual Disabilities Compared With the General Community

Two reviews of obesity research identified a higher prevalence of overweight and obesity among adults with intellectual disability than in the general community (Melville, Hamilton, Hankey, Miller, & Boyle, 2007; Rimmer & Yamaki, 2006). Several studies have supported this conclusion, both for users of formal intellectual and developmental disabilities services and for population samples that include many individuals with intellectual disabil- ity living outside the formal service system (Mel- ville et al., 2008; Yamaki, 2005). However, other studies have found more limited differences (Bhaumik, Watson, Thorp, Tyrer, & McGrother, 2008; Emerson, 2005).

Yamaki (2005) used national population sam- ples from the annual National Health Interview Survey (NHIS) to compare BMI based on self- reported height and weight (for individuals unable to respond, another adult household member could provide the information) of adults with intellectual disability and adults from the general population. The NHIS is a household sample survey of the health status of the U.S. ‘‘noninsti- tutionalized’’ population and includes adults with intellectual disability living with family members or in their own homes but generally excludes persons living in formal service settings. This likely yields more individuals with mild or moderate intellectual disability and fewer comorbid physical, health, and mental health conditions. Yamaki’s operational definition of intellectual disability in- cluded only people who reported a substantial functional limitation and mentioned ‘‘mental retardation’’ as the cause. This definition may not include people who report Down syndrome, autism, cerebral palsy, and other intellectual or developmental disabilities (Hendershot, Larson, Lakin, & Doljanac, 2005).

Compared with the general population, Yamaki (2005) found a higher percentage of adults with intellectual disability in the obese category but no significant overall differences for the overweight category, although men with intellectual disability

had significantly lower prevalence of overweight than men in the general population. In the most recent period examined (1997–2000), 34.6% of adults with intellectual disability were obese com- pared with 20.6% of adults (aged 18–65 years) from the general population, whereas 28.9% (intellectual disability) and 34.1% (general population) were overweight (BMI 5 25.0–30.0). Yamaki’s samples of adults with intellectual disability were moderately sized (range 5 650–1,098), although the most recent sample (1997–2000) of 650 participants yielded relatively large confidence intervals (6 8.0%). Therefore, subgroup analyses were only possible for gender and age group separately, with no examination of race–ethnicity.

Several larger scale and/or population-based studies of BMI have focused on adults with intellectual disability living outside the United States. Overweight and obesity may vary by nation, both for the general population and for those with intellectual disability, with higher prevalence among U.S. individuals (Harris, Rosenberg, Jangda, & Gallagher, 2003; Sassi, Cecchini, & Devant, 2010). Therefore, caution is warranted when reviewing research findings on BMI for persons with intellectual disability from other countries for relevance to U.S. populations.

Emerson (2005; N 5 1,304) found that 14% of disability-accommodation service users in northern England were underweight, 28% over- weight, and 27% obese. Prevalence of obesity among men with intellectual disability did not differ significantly from English men without intellectual disability, except that men with intellectual disability aged 65–74 years had significantly lower obesity rates than men of the same age from the general population. However, women with intellectual disability had higher prevalence of obesity in several age groups and did not differ from women without intellectual disability in other age groups.

Bhaumik et al. (2008; N 5 1,119) examined all individuals on a register of adults with moderate, severe, or profound intellectual disability in a de- fined geographical area in Leicestershire, England. They found that 20.7% of adults with intellectual disability were obese and an additional 28.0% were overweight. The overall intellectual disability sample did not differ significantly from the general popu- lation in England in the prevalence of obesity. Compared with men in the general population (19% obesity prevalence), men with intellectual disability (15% obesity prevalence) had nonsignificantly lower

VOLUME 116, NUMBER 6: 401–418 | NOVEMBER 2011 AJIDD

Obesity among adults R. J. Stancliffe et al.

402 E American Association on Intellectual and Developmental Disabilities

prevalence of obesity. Women with intellectual disability, however, had significantly higher preva- lence of obesity (32%) than women in the general population (23%).

Gender

There is higher prevalence of obesity among women with intellectual disability compared with men with intellectual disability (Bhaumik et al., 2008; Emerson, 2005; Melville et al., 2007, 2008; Robertson et al., 2000; Yamaki, 2005). Melville et al. (2007) concluded that, relative to the higher rate of obesity in women in the general popula- tion, among people with intellectual disability, ‘‘the gender effect is accentuated, placing women with intellectual disabilities at particular risk’’ (p. 225).

Diagnosis and Level of Intellectual Disability

Among adults with intellectual disability, there are important differences in BMI related to diagnosis. Individuals with Down syndrome are more likely to be overweight or obese than other individuals with intellectual disability (Bhaumik et al., 2008; Hove, 2004; Melville et al., 2007, 2008; Rubin, Rimmer, Chicoine, Braddock, & McGuire, 1998; Robertson et al., 2000; Stancliffe et al., in press). Lower prevalence rates of overweight and obesity are evident for adults with cerebral palsy (Bhaumik et al., 2008; Stancliffe et al., in press).

Likewise, level of intellectual disability has been associated with BMI status. Individuals with milder disability have a higher prevalence of obesity, whereas those with more severe disability have a lower rate of obesity but a higher prev- alence of underweight (Emerson, 2005; Hove, 2004; Melville et al., 2007, 2008; Robertson et al., 2000).

Living Arrangements

Living arrangements appear to be related to BMI, with a higher prevalence of obesity evident in less restrictive settings (own home, family home), and lower prevalence in more regulated, fully supervised settings (Melville et al., 2007; Rimmer & Yamaki, 2006). However, not all the studies reporting such findings controlled for differences in personal characteristics between living arrangements. For example, although Lewis,

Lewis, Leake, King, and Lindemann (2002) reported significant differences in level of intel- lectual disability by living arrangements, the lower prevalence of obesity in community group-care facilities may be attributable to the much more severe level of intellectual disability of residents in these settings compared with those living on their own or with family members. When Melville et al. (2008) used multivariate analysis that controlled for level of intellectual disability, they found a significant effect of living arrangements for Scottish women (women living independently were more likely to be obese than those living with family), but no effect for Scottish men. In addition, Melville et al. found no significant multivariate difference by living arrangements for either gender on prevalence of overweight.

Purpose of This Article

The goal of this article is to report on the prevalence of obesity and overweight among adult users of U.S. intellectual disability/developmental disability services in a large sample drawn from the 2008–2009 National Core Indicators (NCI) program and compare these findings to preva- lence data for the general population from Flegal et al.’s (2010) findings from the 2007–2008 National Health and Nutrition Examination Survey (NHANES), with subgroup analysis by age, gender, and race–ethnicity. In addition, we provide descriptive information about BMI of adults with intellectual disability and compare BMI status and obesity prevalence among indi- viduals with different syndromes related to intellectual disability, with different levels of intellectual disability, and with different living arrangements.

Method

Participating States The NCI program is a voluntary collabora-

tion between the National Association of State Directors of Developmental Disabilities Services, the Human Services Research Institute, and state developmental disability agencies of participating states. No NCI data are collected in nonpartici- pating states.

The 8,911 sample members in this study were drawn from all 20 states that participated in the 2008–2009 NCI program and collected con- sumer survey data. Participants were adult users of

VOLUME 116, NUMBER 6: 401–418 | NOVEMBER 2011 AJIDD

Obesity among adults R. J. Stancliffe et al.

E American Association on Intellectual and Developmental Disabilities 403

developmental disabilities services in Alabama, Arkansas, Connecticut, Delaware, Georgia, Illinois, Indiana, Kentucky, Louisiana, Massachusetts, Mis- souri, North Carolina, New Jersey, New York, Ohio, Oklahoma, Pennsylvania, South Carolina, Texas, and Wyoming. Within each participating state, samples were randomly drawn from the state’s population of adults (age $ 18 years) with intellectual disability receiving institutional, com- munity, or home-based services, or some subset of these (a few states’ samples included only recipients of home and community-based services). Sample sizes in participating states ranged from 193 (DE) to 1,502 (NY) and averaged 578.

Instrument Data were collected using the 2008–2009 NCI

Consumer Survey. The 2008–2009 survey was the first version of the NCI survey to obtain data on height and weight, allowing BMI to be calculated. Height and weight data were not measured directly by NCI interviewers but were obtained typically from individual records, setting admin- istrators, or support providers (including family members for participants living with family). These informants provided data on height in feet and inches and data on weight in pounds. These data are reported in the NCI Background section, which requests information on the service user’s personal characteristics, functioning, level of intellectual disability, diagnoses, health, problem behavior, living arrangements, and services. Data in this section are typically obtained from agency records, and it is usually completed by a case manager–service coordinator.

One item asks whether the person has a diagnosis of intellectual disability. The next item asks about the person’s level of intellectual disability (respondents may check one of the following: N/A [not applicable], mild, moderate, severe, profound, unspecified, or unknown). The item that follows asks about a list of 16 other disabilities and diagnoses that are noted on the person’s record (respondents may check all that apply), including autism spectrum disorder, cere- bral palsy, Down syndrome, and Prader-Willi syndrome. The residence-type item provides respondents with 10 response options: specialized institutional facility for persons with intellectual disability/developmental disability, group home, agency-operated apartment, independent home or apartment, parent–relative’s home, foster care or

host home (person lives in home of unrelated, paid caregiver), nursing facility, homeless, other, or ‘‘don’t know.’’ There is also an item on the number of people with disabilities living at the setting, which can be used to cross-check resi- dence type (e.g., an institution is considered to house 16 or more people with a disability). No specific distinction is made between intermediate care facility for people with ‘‘mental retardation’’ (ICFs/MR) and settings with other funding or regulatory arrangements, in that ICFs/MR can be classified as institutions ($16 residents) or group homes (#15 residents), but group homes also include non-ICF/MR settings.

Interviewer training. To ensure that all inter- viewers received consistent training, the NCI Consumer Survey protocol is supported by a training program for interviewers, including training manuals, presentation slides, training videos, scripts for scheduling interviews, and lists of frequently asked questions. The training includes question-by-question review of the sur- vey tool.

Reliability. Multiple tests of the NCI instruments have yielded interrater agreement of 92%–93%, and a single examination of test–retest reliability resulted in 80% agreement (Smith & Ashbaugh, 2001). How- ever, no item testing was done on the specific height and weight variables.

Participants The total 2008–2009 NCI sample consisted

of 11,569 individual users of adult intellectual disability/developmental disability services from 20 states. We excluded 99 people whose age was missing and another 208 sample members aged 18 or 19 years, because our general population comparison sample only included adults aged 20 or older (Flegal et al., 2010). Because we also wanted to compare our sample with the compar- ison sample on gender, race, and age, we excluded individuals with missing data on these variables. In addition, we excluded 421 individuals whose height was missing and 13 adults with recorded heights of 36 inches or less or 84 inches or more. Such listed heights were, of course, possible but were notable outliers, and we had no means to follow up on their accuracy. Last, we omitted individuals without an intellectual disability diagnosis because our focus was on BMI in adults with intellectual disability. This selection process yielded a final sample of 8,911 individuals from

VOLUME 116, NUMBER 6: 401–418 | NOVEMBER 2011 AJIDD

Obesity among adults R. J. Stancliffe et al.

404 E American Association on Intellectual and Developmental Disabilities

20 states, with an average age of 43.48 years (range 5 20–93 years).

The U.S. general population comparison data, including breakdowns by gender and race, were drawn from analyses of the NHANES by Flegal et al. (2010). To enable close comparison, the NCI sample was broken down according to the same age groups and as similar as possible racial groups. Individuals were grouped by age as follows: 20–39 years, 40–59 years, and 60 years or older. Race and ethnicity were classified as non- Hispanic White, non-Hispanic Black, and His- panic/other. The first two of these race categories were identical to the comparison group from Flegal et al. Table 1 shows sample numbers by racial group, gender, and age group. Information about participant numbers by level of intellectual disability and living arrangements is shown in the results section.

Results

Overweight and Obesity Raw data were gathered on height in feet and

inches and weight in pounds, not in metric units, because these were the standard units reported in the individuals’ health records. These data were used to calculate BMI using the following formula:

BMI~ body mass lbð Þx 703

height ftð Þð Þ2 :

BMI was classified as follows: (a) underweight: BMI , 18.50; (b) normal weight: BMI 5 18.50–24.99; (c)

overweight: BMI 5 25.00–29.99; (d) obese: BMI 5 $30.00; Grade 2 obesity: BMI $ 35.00; Grade 3 obesity: BMI $ 40.00 (World Health Organization [WHO] Expert Committee on Physical Status, 1995).

Comparison with the general population. We calculated the prevalence of overweight and obesity by race, age group, and gender (Table 2). We used Flegal et al.’s (2010, Table 2) analysis of NHANES data as the basis for comparison be- tween persons with intellectual disability (Table 2) and the general population (for those $ 20 years old). Nonoverlap of the 95% confidence intervals between groups was considered to be a significant difference. To assist with comparison, selected groupings of Flegal et al.’s data (all people, all men, all women) are reproduced in Table 2 along with the corresponding groupings for people with intellectual disability drawn from the NCI sample. Readers should consult Flegal et al.’s Table 2 directly for more detailed comparisons for specific age and gender groups. Because the Hispanic and Mexican American samples were constituted differently in the general population data (Flegal et al.) than in the NCI data, we present the data for the Hispanic/other group without a general population comparison.

On most comparisons, our sample of adult service users with intellectual disability from 20 states did not differ from the nationally represen- tative sample of the U.S. population. Of 27 possible comparisons for obesity prevalence, only 4 were significant (denoted in Table 2 by an asterisk to indicate that the NCI subgroup mean differed significantly from the corresponding subgroup mean in Flegal et al.’s, 2010, Table 2).

Table 1. Number of Sample Members with Intellectual Disability by Racial Group, Gender, and Age

Gender/age group

Total

(N 5 8,911)

Non-Hispanic White

(n 5 6,488)

Non-Hispanic Black

(n 5 1,706)

Hispanic/other

(n 5 717)

Men

20–39 2,152 1,394 499 259

40–59 2,248 1,665 444 139

60+ 605 496 78 31 All men 5,005 3,555 1,021 429

Women

20–39 1,553 1,095 305 153

40–59 1,751 1,319 329 103

60+ 602 519 51 32 All women 3,906 2,933 685 288

VOLUME 116, NUMBER 6: 401–418 | NOVEMBER 2011 AJIDD

Obesity among adults R. J. Stancliffe et al.

E American Association on Intellectual and Developmental Disabilities 405

T a b

le 2 .

P er

ce n

ta ge

o f

A d

u lt

s W

it h

In te

ll ec

tu al

D is

ab il it

ie s

an d

a C

o m

p ar

is o

n S am

p le

F ro

m th

e G

en er

al P o

p u

la ti

o n

a W

h o

W er

e O

b es

e o

r O

v er

w ei

gh t

b y

R ac

e, A

ge ,

an d

G en

d er

S a m

p le

g e n

d e r/

a g

e g

ro u

p

P e rc

e n

ta g

e o

f a d

u lt

s (9

5 %

co n

fi d

e n

ce in

te rv

a l)

A ll

(N C

I (N

5 8 ,9

1 1 )

N o

n -H

is p

a n

ic W

h it

e

(N C

I n

5 6 ,4

8 8 )

N o

n -H

is p

a n

ic B

la ck

(N C

I n

5 1 ,7

0 6 )

H is

p a n

ic /O

th e

r2

(N C

I n

5 7 1 7 )

O b

e se

: B

M I

$ 3 0

T o

ta l

U .S

. sa

m p

le 3 3 .8

(3 1 .6

– 3

6 .0

) 3 2 .8

(2 8 .9

– 3 5

.9 )

4 4 .1

(4 0 .0

– 4 8

.2 )

N C

I sa

m p

le 3 3 .6

(3 2 .6

– 3

4 .6

) 3 3 .0

(3 1 .9

– 3 4

.2 )A

3 5 .8

(3 3 .5

– 3 8

.0 )*

A 3 3 .2

(2 9 .7

– 3 6 .7

)A

M e n

U .S

. sa

m p

le :

a ll

3 2 .2

(2 9 .5

– 3

5 .0

) 3 1 .9

(2 8 .1

– 3 5

.7 )

3 7 .3

(3 2 .3

– 4 2

.4 )

N C

I sa

m p

le :

a ll

2 9 .4

(2 8 .2

– 3

0 .7

)a 2 9 .6

(2 8 .1

– 3 1

.1 )c

B 2 8 .6

(2 5 .8

– 3 1

.4 )*

e B

2 9 .6

(2 5 .3

– 3 3 .9

)g B

N C

I: 2 0 – 3 9

3 0 .8

(2 8 .9

– 3

2 .8

)a 3 0 .7

(2 8 .3

– 3 3

.1 )c

C 3 0 .1

(2 6 .0

– 3 4

.1 )e

C 3 2 .8

(2 7 .1

– 3 8 .6

)g C

N C

I: 4 0 – 5 9

2 9 .7

(2 7 .8

– 3

1 .6

)a 3 0 .5

(2 8 .2

– 3 2

.7 )c

D 2 7 .9

(2 3 .7

– 3 2

.1 )e

D 2 6 .6

(1 9 .2

– 3 4 .1

)g D

N C

I: 6 0

+ 2 3 .3

(1 9 .9

– 2

6 .7

)* a

2 3 .8

(2 0 .0

– 2 7

.6 )*

c E

2 3 .1

(1 3 .5

– 3 2

.6 )e

E 1 6 .1

(2 .4

– 2 9 .8

)g E

W o

m e n

U .S

. sa

m p

le :

a ll

3 5 .5

(3 3 .2

– 3

7 .7

) 3 3 .0

(2 9 .3

– 3 6

.6 )

4 9 .6

(4 5 .5

– 5 3

.7 )

N C

I sa

m p

le :

a ll

3 8 .9

(3 7 .4

– 4

0 .4

)b 3 7 .2

(3 5 .5

– 3 9

.0 )d

F 4 6 .4

(4 2 .7

– 5 0

.2 )f

G 3 8 .5

(3 2 .9

– 4 4 .2

)g F G

N C

I: 2 0 – 3 9

4 1 .1

(3 8 .7

– 4

3 .6

)b 3 8 .2

(3 5 .3

– 4 1

.1 )d

H 5 1 .5

(4 5 .8

– 5 7

.1 )f

I 4 1 .8

(3 3 .9

– 4 9 .7

)g H

I

N C

I: 4 0 – 5 9

3 9 .5

(3 7 .2

– 4

1 .8

)b 3 8 .9

(3 6 .3

– 4 1

.5 )d

J 4 1 .9

(3 6 .6

– 4 7

.3 )f

J 3 9 .8

(3 0 .2

– 4 9 .4

)g J

N C

I: 6 0

+ 3 1 .4

(2 7 .7

– 3

5 .1

)b 3 0 .8

(2 6 .8

– 3 4

.8 )c

K 4 5 .1

(3 1 .0

– 5 9

.2 )e

K 1 8 .8

(4 .5

– 3 3 .1

)g K

O v e rw

e ig

h t

o r

o b

e se

: B

M I

$ 2 5

T o

ta l

U .S

. sa

m p

le :

a ll

6 8 .0

(6 6 .3

– 6

9 .8

) 6 6 .7

(6 4 .1

– 6 9

.3 )

7 3 .8

(7 1 .3

– 7 6

.3 )

N C

I sa

m p

le :

a ll

6 2 .2

(6 1 .2

– 6

3 .3

)* 6 2 .0

(6 0 .8

– 6 3

.2 )*

L 6 3 .8

(6 1 .5

– 6 6

.1 )*

L 6 0 .7

(5 7 .1

– 6 4 .3

)L

M e n

U .S

. sa

m p

le :

a ll

7 2 .3

(7 0 .4

– 7

4 .1

) 7 2 .6

(6 9 .9

– 7 5

.3 )

6 8 .5

(6 5 .2

– 7 1

.8 )

N C

I sa

m p

le :

a ll

6 0 .5

(5 9 .1

– 6

1 .8

)* h

6 0 .7

(5 9 .1

– 6 2

.3 )*

jM 6 0 .3

(5 7 .3

– 6 3

.3 )*

k M

5 9 .0

(5 4 .3

– 6 3 .7

)m M

N C

I: 2 0 – 3 9

5 7 .7

(5 5 .6

– 5

9 .8

)* h

5 8 .0

(5 5 .6

– 6 0

.6 )*

jN 5 7 .0

(5 2 .6

– 6 1

.3 )k

N 5 7 .9

(5 1 .9

– 6 4 .0

)m N

N C

I: 4 0 – 5 9

6 3 .9

(6 1 .9

– 6

5 .9

)* h

6 3 .8

(6 1 .4

– 6 6

.0 )*

jO 6 5 .5

(6 1 .1

– 7 0

.0 )k

O 6 1 .2

(5 3 .0

– 6 9 .4

)m O

N C

I: 6 0

+ 5 7 .4

(5 3 .4

– 6

1 .3

)* h

5 8 .1

(5 3 .7

– 6 2

.4 )*

jP 5 2 .6

(4 1 .2

– 6 3

.9 )*

k P

5 8 .1

(3 9 .7

– 7 6 .5

)m P

(T a b

le 2

c o n

ti n

u e d

)

VOLUME 116, NUMBER 6: 401–418 | NOVEMBER 2011 AJIDD

Obesity among adults R. J. Stancliffe et al.

406 E American Association on Intellectual and Developmental Disabilities

In each case, the group with intellectual disability had lower prevalence of obesity.

For the overweight and obesity prevalence data (lower half of Table 2), 15 of 27 (56%) comparisons were significant. Here too, the group with intellec- tual disability had lower prevalence in every case. Ten of the 12 (83%) comparisons involving men were significant, suggesting a lower prevalence of overweight and obesity among men with intellec- tual disability than men in the general U.S. population. By contrast, only 2 of the 12 (17%) comparisons for women were significant, suggest- ing that, for women, there is a similar prevalence of overweight and obesity to the general population.

Gender comparisons. Among the sample of people with intellectual disability, men were significantly less obese than women (significant pairwise gender comparisons denoted in Table 2 by means with different lowercase superscript letters when comparing the equivalent male and female subgroup within columns). This was true for the sample overall (all men vs. all women) and for non-Hispanic White and non-Hispanic Black men but not for men in the Hispanic/other group. These significant gender differences in obesity prevalence were also true for within-race compar- isons between men and women in the 20–39-year- old and 40–59-year-old age groups for White and Black participants. The absence of significant gender differences for the 60+ age group may have been due, in part, to its much smaller sample size.

Among adults with intellectual disability, gender differences were less evident for the combined overweight and obesity groups. There was a significantly lower prevalence of overweight and obesity when comparing all men (60.5%) with all women (64.5%), as well as for all Black men with all Black women, and for all men and Black men among the 20–39-year-old age group. These differences appear to have been driven by the higher obesity prevalence in women. Indeed, taking overweight (25.0 # BMI , 30.0) prevalence only, 31.0% of all men and 25.6% of all women were overweight.

Comparisons by race–ethnic group. There were no significant differences by race–ethnic group in the prevalence of obesity or of combined overweight and obesity among the sample as a whole or among men with intellectual disability. However, signifi- cantly more Black women were obese than White women (significant race–ethnic group comparisons denoted in Table 2 by means with different upper- case superscript letters within rows). When brokenT

a b

le 2 .

C o

n ti

n u

ed .

S a m

p le

g e n

d e r/

a g

e g

ro u

p

P e rc

e n

ta g

e o

f a d

u lt

s (9

5 %

co n

fi d

e n

ce in

te rv

a l)

A ll

(N C

I (N

5 8 ,9

1 1 )

N o

n -H

is p

a n

ic W

h it

e

(N C

I n

5 6 ,4

8 8 )

N o

n -H

is p

a n

ic B

la ck

(N C

I n

5 1 ,7

0 6 )

H is

p a n

ic /O

th e

r2

(N C

I n

<p