The prevalence of overweight and obesity in the United States has increased at an alarming rate over the past several decades [
Acknowledging that few studies have explored race-specific trends in overweight/obesity according to levels of educational attainment over time [
To gain a better understanding of current temporal trends related to the influence of educational attainment on overweight/obesity disparities while addressing important gaps in the literature, we used a considerably large, nationally representative sample of the noninstitutionalized US black and white population. We hypothesized that (1) the prevalence of overweight and obesity will have reached a peak among blacks over time, with whites steadily catching up and (2) the racial disparity in overweight/obesity will be wider in groups with higher compared to lower educational attainment, especially among women.
We analyzed data from NHIS—a series of cross-sectional, nationally representative surveys which used a three-stage stratified cluster probability sampling design to conduct in-person interviews in samples of noninstitutionalized US civilian households. A complete description of NHIS procedures is available elsewhere [
Participants included self-reported non-hispanic white or non-hispanic black (henceforth, white and black) adults aged 25 through 75 years. Participants were excluded if they (1) were born outside the US; (2) reported having a history of cancer and/or heart disease; (3) were pregnant; (4) had missing data on height, weight, educational attainment, or smoking status; or (5) had an extreme body mass index (BMI)—that is, either <15 or >55 kg/m2. Our final sample comprised of 174,228 adults (Figure
Study flow diagram.
Self-reported height and weight were used to calculate BMI (kg/m2). Obesity was defined as BMI ≥30 kg/m2, overweight as 25.0–29.9 kg/m2, normal weight as 18.5–24.9 kg/m2, and underweight as BMI <18.5 kg/m2.
Educational attainment was categorized as less than high school (<HS) (no high school diploma), high school (HS) (high school or general equivalency diploma), and greater than high school (>HS) (any education beyond high school).
Smoking status was categorized as “ever” or “never.” Lifetime alcohol drinking status was assessed and categorized as either “ever” or “never.” Leisure-time physical activity was categorized as none, low, or high based on the participant’s answer to the following questions: (1) “How often do you do light or moderate leisure-time physical activities for at least 10 minutes that cause only light sweating or a slight to moderate increase in breathing or heart rate?” and (2) “How often do you do vigorous physical activities for at least 10 minutes that cause heavy sweating or a large increase in breathing or heart rate?” Individuals who answered “never” or “unable to do this type activity” were classified as “none.” Those engaging in at least some level of activity and providing a specific number of activity bouts were dichotomized at the midpoint of these bouts into “low” or “high.” Marital status was categorized as married/living with partner, divorced/separated/widowed, or never married, and regions of the country as South, Midwest, Northeast, and West.
We used 12 years (1997–2008) of NHIS data merged by the Integrated Health Interview Series [
Continuous variables were expressed as means ± standard errors (SE), whereas categorical variables were presented as absolute values with corresponding percentages. To test for differences in prespecified sociodemographic, clinical, and behavioral characteristics between whites and blacks and by obesity status, we used the Rao-Scott second-order corrected Pearson statistic [
Poisson regression models were used to estimate prevalence ratios and corresponding 95% confidence intervals adjusted for age (in 4 categories: 25–34, 35–49, 50–64, and 65–75 years), marital status, smoking status, alcohol consumption, leisure-time physical activity level, and region of the country [
Differences in linear trends in mean BMI from 1997 to 2008 between blacks and whites within each educational attainment category were formally tested (at the
Sociodemographic, clinical, and behavioral characteristics of the final sample of 174,228 NHIS study participants are shown by race and educational attainment in Table
Sociodemographic, health behavior, and clinical characteristics of National Health Interview Survey (NHIS) participants, according to educational attainment and race/ethnicity, 1997–2008 (
<HS | HS | >HS | Combined education | ||||||
---|---|---|---|---|---|---|---|---|---|
White | Black | White | Black | White | Black | White | Black | Total | |
Sample size, |
11,987 (76) | 5,617 (24) | 45,174 (85) | 11,316 (15) | 85,189 (89) | 14,945 (11) | 142,350 (87) | 31,878 (13) | 174,228 |
Age | |||||||||
Mean, year ± SE | 50.9 ± 0.15 | 50.5 ± 0.27 | 47.4 ± 0.08 | 43.9 ± 0.15 | 44.6 ± 0.07 | 42.1 ± 0.13 | 46.0 ± 0.05 | 44.1 ± 0.11 | 45.8 ± 0.05 |
Age group—% | |||||||||
25–34 | 18 | 18 | 20 | 27 | 25 | 31 | 23 | 28 | 23 |
35–49 | 30 | 30 | 39 | 43 | 42 | 43 | 40 | 41 | 40 |
50–64 | 29 | 31 | 28 | 23 | 26 | 21 | 27 | 23 | 27 |
65–75 | 23 | 21 | 13 | 7 | 7 | 5 | 10 | 8 | 10 |
Male | 53 | 46 | 51 | 49 | 51 | 43 | 51 | 46 | 51 |
Marital status | |||||||||
Married/living w/partner | 59 | 33 | 68 | 41 | 69 | 46 | 68 | 42 | 65 |
Divorced/separated/widowed | 28 | 36 | 20 | 27 | 16 | 25 | 18 | 27 | 19 |
Never married | 13 | 31 | 12 | 32 | 15 | 29 | 14 | 31 | 16 |
Health behaviors | |||||||||
Ever smoker (yes) | 68 | 58 | 57 | 44 | 42 | 34 | 49 | 41 | 46 |
Alcohol consumer | |||||||||
Never | 46 | 50 | 36 | 48 | 24 | 47 | 30 | 48 | 32 |
Ever | 54 | 50 | 64 | 52 | 76 | 53 | 70 | 52 | 68 |
Leisure-time physical activity | |||||||||
Never/unable | 59 | 67 | 42 | 53 | 24 | 36 | 33 | 47 | 34 |
Low | 19 | 17 | 29 | 25 | 38 | 35 | 34 | 29 | 33 |
High | 22 | 16 | 29 | 22 | 38 | 29 | 34 | 24 | 32 |
Clinical characteristics | |||||||||
Mean BMI (kg/m2), men | 27.9 ± 0.09 | 27.8 ± 0.13 | 28.0 ± 0.04 | 28.2 ± 0.09 | 27.6 ± 0.03 | 28.7 ± 0.08 | 27.7 ± 0.02 | 28.4 ± 0.06 | 27.8 ± 0.02 |
Mean BMI (kg/m2), women | 27.7 ± 0.10 | 30.6 ± 0.15 | 27.0 ± 0.05 | 29.6 ± 0.10 | 25.9 ± 0.04 | 29.1 ± 0.09 | 26.4 ± 0.03 | 29.5 ± 0.07 | 26.8 ± 0.03 |
Overweight/obese (yes) | 66 | 73 | 65 | 73 | 59 | 72 | 62 | 72 | 63 |
Obesity (yes) | 31 | 40 | 28 | 36 | 22 | 35 | 25 | 36 | 26 |
Health conditions | |||||||||
Hypertension (yes) | 34 | 48 | 26 | 32 | 19 | 28 | 23 | 33 | 24 |
Diabetes (yes) | 11 | 16 | 6 | 9 | 4 | 8 | 5 | 9 | 6 |
General health status | |||||||||
Excellent/very good | 39 | 32 | 61 | 50 | 77 | 63 | 69 | 54 | 66 |
Good | 34 | 32 | 29 | 33 | 18 | 27 | 23 | 30 | 24 |
Fair/poor | 27 | 36 | 10 | 17 | 5 | 10 | 9 | 17 | 10 |
Region of country | |||||||||
Northeast | 15 | 12 | 20 | 13 | 19 | 13 | 19 | 13 | 18 |
Midwest | 24 | 17 | 30 | 19 | 28 | 20 | 28 | 19 | 27 |
South | 47 | 67 | 35 | 61 | 33 | 57 | 35 | 60 | 38 |
West | 14 | 4 | 15 | 7 | 20 | 10 | 18 | 8 | 17 |
Weighted estimates; mean ± SE or (%); SE: standard error; <HS: less than high school, HS: high school, >HS: greater than high school.
Blacks were slightly younger than whites, less likely to be married, and more likely to reside in the Southern region of the United States. Blacks were more likely to have less than a high school education, to be obese, to report having hypertension, to report never consuming alcohol, and to report having no leisure-time physical activity. All participant characteristics had less than 10% missing values. We compared participants with complete data versus their counterparts with missing data and found no significant differences in age, sex, race, health status, poverty status, or household size.
From 1997 to 2008, BMI increased by at least 1 kg/m2 in all race-sex-education groups (except black men with less than high school education), and mean BMI appeared to increase at a faster pace among whites compared to blacks (Figure
Smoothed trends in mean body mass index among (a) women and (b) men.
While mean BMIs were different by race (especially among women) the unadjusted slope of BMI increase was significantly different for men (
Slope differences for BMI and survey year between Blacks compared to Whites among men and women, overall and by educational attainment, 1997–2008.
|
95% confidence interval |
|
|
95% confidence interval |
|
|
---|---|---|---|---|---|---|
Men | ||||||
<HS |
|
(−0.20641–−0.01457) | 0.02 |
|
(−0.21388–−0.02291) | 0.02 |
HS |
|
(−0.11559–0.00399) | 0.07 |
|
(−0.10294–0.01287) | 0.13 |
>HS |
|
(−0.06770–0.04205) | 0.65 |
|
(−0.06803–0.04169) | 0.64 |
Combined education | 0.00020 | (0.00012–0.00027) | <0.001 | 0.00030 | (0.00023–0.00037) | <0.001 |
Women | ||||||
<HS |
|
(−0.13259–0.09085) | 0.71 |
|
(−0.14755–0.07082) | 0.49 |
HS |
|
(−0.06656–0.05762) | 0.89 |
|
(−0.06251–0.06162) | 0.99 |
>HS |
|
(−0.03016–0.06961) | 0.44 |
|
(−0.04473–0.05174) | 0.89 |
Combined education |
|
(−0.00044–−0.00029) | <0.001 |
|
(−0.00010–0.00004) | 0.44 |
Adjusted model: age (4 categories), marital status, smoking status, leisure-time physical activity, alcohol consumption, poor income, region of country, and self-reported general health status;
Figure
Adjusted prevalence ratios on a log scale for overweight/obesity for blacks compared to whites by sex and educational attainment in 1997 to 2000, 2001 to 2004, and 2005 to 2008.
Our analysis of overweight/obesity prevalence trends by sex, race, and educational attainment among US-born Non-Hispanic black and white adults showed that BMI has increased steadily from 1997 to 2008 in all race-sex and education groups, with the exception of black men with less than a high school education, whose prevalence of obesity appeared steady. The racial disparity in overweight/obesity prevalence remained largely proportional over time for each respective education group among women, but the disparity differed by level of educational attainment among men. As a result of the rate of BMI increase being lowest among black men with less than a high school education, there is currently little difference in BMI between black and white men. Blacks (especially women) had a consistently higher BMI than their white counterparts. Although white women remained the leanest group throughout the study period, their mean BMI was above 26 kg/m2 at the end of the study.
The Coronary Artery Risk Development in Young Adults (CARDIA) study included participants (5,115 black and white men and women) with ages that ranged from 18 to 30 years and found an inverse, cross-sectional association of education with obesity among white women, a positive association among black men, and no significant relationship among both white men and black women [
Using NHANES data pooled from 1999 to 2000, overall, persons in the United States with less than a high school education had a higher prevalence of obesity than their counterparts with more education, with the exception of black women with less than a high school education who had the lowest obesity prevalence compared to those with a higher level educational attainment [
We identified two studies that investigated time trends in racial differences in obesity prevalence by SES [
Another study investigated trends in obesity over time using NHANES data from 1999 to 2004 and found that obesity increased in adults at all education levels [
Although few studies with nationally representative data report small differences in weight and height self-reporting error between blacks and whites of both genders [
Suggested mechanisms for the generally inverse education-overweight/obesity association include differences in healthy lifestyles and social-psychological resources, in addition to work and economic conditions [
Although it has been widely accepted that low-SES US groups are at increased risk of obesity [
There are several limitations of our study that deserve to be mentioned. First, our data are based on self-report, and thus differential misclassification by race in ascertaining height and weight to estimate BMI is possible. Second, the three educational levels were fairly broad, and it is likely that blacks are closer to the lower cut-off of each education level than whites, which may result in residual confounding. Additionally, even at the same levels, education, as a marker of SES, may not have the same social and health benefits or construct validity across racial/ethnic groups [
Important to our relationships of interest, we were also unable (like many studies) to assess education quality versus attainment, adjust for finer categories of smoking status, which is associated with lower BMI, or discern cohabiting couples from those who were married or single. Despite these limitations, our study has several strengths. First, the sample size was large, allowing stratification by race and educational attainment. Second, we had a relatively large black population, which affords more robust estimates than those from previous studies as this is the largest sample of the US population. In addition, we used a nationally representative sample of the USA.
These data help enable planners to develop more effective public health strategies and direct resources to subpopulations with exceptionally high (or increasing) obesity for targeted intervention and in-depth research. As most race-sex-education groups have been affected by the obesity epidemic, a growing consensus of stakeholders agrees that population-based policies and programs emphasizing environmental changes are most likely to be successful in addressing the obesity epidemic.
This study underscores substantial and complex differences in obesity prevalence by education (especially among women), which have persisted over time. Black women with greater than a high school education had substantially higher mean BMIs than even white women with less than a high school education. Our study suggests that mean BMI appears to be increasing at a faster pace among whites than blacks and racial disparities in overweight/obesity trends and prevalence were more prominent among more educated individuals as compared to their less-educated counterparts. Higher education does not appear protective against the obesity epidemic nor racial/ethnic disparities in overweight/obesity.
The authors have no conflict of interests to declare.
Dr. Brancati was supported by a Grant from NIDDK (K24 DK62222). Drs. Yeh, Wang, and Brancati were supported by a Diabetes Research and Training Center Grant, also from NIDDK (P60 DK079637). Dr. Thorpe is supported by a NIH P60 Center of Excellence NIMHD Center Grant (P60MD000214). Dr. Jackson was supported by a C. Sylvia and Eddie C. Brown Scholarship from the Johns Hopkins Bloomberg School of Public Health and the TREC Grant (1U54CA155626-01).