3% WWNHNUNNHUHWWNWW‘HflWll‘lHlI

Tints." ‘fiie‘

fill .“PRARY
Michigan State
University

This is to certify that the
thesis entitled

ADULT AND PRE—ADULT SOCIOECONOMIC IN DICES AND
PRE-PREGNANCY OVERWEIGHT AND OBESITY

presented by
Debora S. Tauiliili

has been accepted towards fulfillment
of he requirements for the

MS. degree in Epidemiology

 

 

 
   

s r’s Signature

[412%th Ag 200?

Date

MSU is an aflMnetlvo—ecflon, equal-opportunity employer

 

—.—.-.----u--u-u-u-o---.-*--v---I-n-.--.-.- _... - _t_.—4—.-.' u.-.—.--u--r---r-'-

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KLIProlecc8Pres/ClRC/DaleDue indd

ADULT AND PRE-ADULT SOCIOECONOMIC INDICES AND PRE-PREGNANCY
OVERWEIGHT AND OBESITY

By

Debora S. Tauiliili

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Epidemiology

2008

ABSTRACT

ADULT AND PRE-ADULT SOCIOECONOMIC INDICES AND PRE-PREGNAN CY
OVERWEIGHT AND OBESITY

By
Debora S. Tauiliili

In light of the overall increase in overweight and obesity in the general population,
it is not surprising that the prevalenCe of pre-pregnancy overweight and obesity have also
dramatically increased over the past decade. The main objective of this study was to
examine associations between adult and pre-adult socioeconomic factors, SES mobility,
and pre-pregnancy overweight and obesity. Pre—pregnancy overweight and obesity were
calculated based on self-reported height and weight in a sample of 2876 women, aged 15-
47 years old who participated in the prospective Pregnancy Outcomes and Community
Health (POUCH) study between 1998 and 2004. In White/other women: low SES in pre-
adulthood was associated with pre-pregnancy obesity, AOR (adjusted OR) = 1.9 (95%
CI=1.4-2.7) and low SES in adulthood was associated with pre—pregnancy overweight,
AOR = 1.7 (95% CI=l.2-2.4). In both White/other and Black women: low SES in
adulthood was inversely associated with pre-pregnancy obesity, AOR = 3.4 (95%
CI=2.4-4.8) and 2.8 (95% CI=1.4-5.9), respectively. Social mobility also had an effect
on risk of pre-pregnancy overweight and obesity and results varied by race/ethnicity. The
findings suggest that ongoing public health interventions are needed to continue
encouraging women to adopt or maintain healthy behaviors before pregnancy with a

particular focus on the poor and minorities.

ACKNOWLEDGMENTS

I would like to thank my thesis committee, Drs. Claudia Holzman (Chair), Ellen
Velie, and Hwan Chung for assistance in the development of this work. I am
tremendously gratefiil to Dr. Holzman for her intellectual mentoring and guidance, and
for the sensitivity and understanding displayed towards me when family obligations had
to take precedence over my thesis work. She also allowed me to use data from her
POUCH study which pamitted the testing of very interesting hypotheses. In addition, I
want to extent my sincere gratitude to members of the POUCH research team for their
support throughout my education.

There are many individuals who positively influenced my graduate education in
epidemiology. They are too many to list but among these I would like to highlight the
unconditional support given to me by Dr. Kris Siefert.

Finally, I want to thank my husband and children for their endless support and

patience, for without them this thesis would not have been possible.

iii

TABLE OF CONTENTS

LIST OF TABLES ............................................................... vi
LIST OF FIGURES .............................................................. vii
INTRODUCTION ............................................................... 1
BACKGROUND ................................................................. 4
Obesity and SES .............................................................. 4
Socioeconomic Measures: A Review .................................... 5
Education and Obesity ................................................ 5
Occupation and Obesity ............................................... 6
Income and Obesity ................................................... 7
Other SES Indicators and Obesity ................................... 8
SES Mobility and Obesity ............................................ 8
Pre-pregnancy Overweight, Obesity and SES .......................... 10
Theoretical Framework: The Life Course Perspective ............... 11
Research Question and Hypotheses ...................................... 12
METHODS ....................................................................... 14
Study Population ............................................................ 14
Measures ..................................................................... 15
Dependent Variable — Pre-pregnancy BMI Categories .......... 15
Independent Variables ................................................ 15
Adult SES .......................................................... l6
Pre-adult SES .................................................... 19
SES Mobility ....................................................... 20
Race/ethnicity .................................................... 20
Covcm'ates ........................................................ 21
Analytic Strategy ........................................................... 23
RESULTS ......................................................................... 25
Socioeconomic status over the Life Course ............................ 25
Pre-pregnancy BMI Categories .......................................... 25
Adult SES Indices and Pre-pregmncy BMI Categories ............... 26
Pre-adult SES Indices and I’m-pregnancy BMI Categories .......... 26
Adult and Pre-adult Composite SES Measures ........................ 27
Adult and Pre-adult Composite SES Measures and Pro-pregnancy
BMI Categories ............................................................. 28
Life Course SES, SES Mobility and Pre—pregnancy BMI
Categories ................................................................... 29
DISCUSSION ................................................................... 31

iv

Pre-adult, Adult-SES and Pre-pregnancy BMI Categories .......... 31
SES Over the Life Course and Pre-pregnancy BMI Categories. . .. 34

Study Limitations ......................................................... 36
Implications for Public Health .......................................... 38
REFERENCES .................................................................. 53

LIST OF TABLES

Table 1. Selected demographic characteristics, POUCH study 1998-

Table 2. Adult SES indices and pre-pregnancy BMI categories,
POUCH study, 1998-2004 ...................................................... 41

Table 3. Unadjusted and adjusted risk of pre-pregnancy overweight and
obesity by adult SES indices, POUCH study, 1998-2004 .................. 42

Table 4. Pre-adult SES indices and pre-pregnancy BMI categories,
POUCH study, 1998-2004 ...................................................... 43

Table 5. Unadjusted and adjusted risk of pre-pregnancy overweight and
obesity by pre-adult SES indices, POUCH study, 1998-2004 ............. 44

Table 6. Distribution of pre-pregnancy BMI by categories of high and
low adult & pre-adult SES indices, POUCH study, 1998-2004 ........... 45

Table 7. Risk of pre-pregnancy overweight and obesity by pre-adult
SES composite index and adult SES composite index for Black women
(N=675), POUCH study, 1998-2004 .......................................... 46

Table 8. Risk of pre-pregnancy overweight and obesity by pre-adult
SES composite index and adult SES composite index for White/other
women (N=2124), POUCH study, 1998-2004 ............................... 47

Table 9. Risk of pre-pregnancy overweight and obesity by SES over
the life course for Black women (N=642), POUCH study, 1998-
2004 ................................................................................ 48

Table 10. Risk of pre-pregnancy overweight and obesity by SES
mobility over the life course for Black women, POUCH study, 1998-
2004 ................................................................................ 49

Table 11. Risk of pre-pregnancy overweight and obesity by SES over
the life course for White/other women (N=2054), POUCH study, 1998-
2004 ................................................................................ 50

Table 12. Risk of pre-pregnancy overweight and obesity by SES
mobility over the life course for White/other women, POUCH study,
1998-2004 ......................................................................... 51

LIST OF FIGURES

Figure 1. Predicted probabilities of obesity reflecting the effect of the
interaction of race/ethnicity with the three-category ‘Adult SES’

variable ........................................................................... 52

vii

INTRODUCTION

Maternal pro-pregnancy overweight and obesity have been associated with
multiple complications both in the mother and her offspring. Pre-pregnancy overweight
and obesity have been shown to increase mother’s risk for chronic hypertension,
gestational hypertension, (Bodnar, Catov, Klebanofl‘, et al. 2007; Samuels-Kalow, Funai,
Buhimschi, et al. 2007) and pre-eclampsia (Baeten, Bukusi, Lambe, et al. 2001; Bodnar,
Ness, Markovic, et al. 2005; Bodnar, Catov, Klebanofl‘, et al. 2007; Samuels-Kalow,
Funai, Buhimschi, et al. 2007). In at least one report, women who experienced
hypertensive disorders during pregnancy and were either overweight or obese prior to
being pregnant were at increased risk of dying during the antenatal or postpartum period
(Samuels-Kalow, Funai, Buhimschi, et al. 2007). Pre-pregnancy overweight and obesity
have also been linked to a higher risk of developing gestational diabetes (Baeten, Bukusi,
Lambe, et al. 2001). Gestational diabetesincreases the risk of having a large-for-
gestational age (LGA) fetus, which in turn can lead to a more difficult labor and delivery
(Casey, Lucas, Mcintire, et al. 1997). Gestational diabetes has also been linked to
congenital central nervous system defects (Anderson, Waller, Canfield, et al. 2005), other
types of defects (Waller, Shaw, Rasmussen, et al. 2007), and development of type II
diabetes (Dabelea, Hanson, Lindsay, et al. 2000) overweight, and obesity in offspring
(Plagemann, Harder, Kohlhofl‘ , et al. 1997, Dabelea, Hanson, Lindsay, et al. 2000). Pre-
pregnancy overweight and obesity have also been associated with increased risk of
cesarean deliveries, irrespective of gestational diabetes (Baeten, Bukusi, Lambe, et al.

2001).

Despite society’s knowledge of the problems associated with overweight and
obesity, the prevalence of overweight and obesity have dramatically increased over the
past several decades among women, men, and children in the United States (U .S.) (Flegal,
Carroll, Kuczrnarski, et al. 1998; Flegal, Carroll, Ogden, et al. 2002; Caban, Lee, Fleming,
et al. 2005; Wang & Beydoun 2007). Mean body mass index (BMI) for women went
from 25.3 in 1976-1980 to 28.2 kg/m2 in 1999-2000. Using mete data, the authors
projected that in the next 10 years, women’s mean BMI will be at the level of obesity, a
substantial shift from the present mean, which is in the overweight range (Wang &
Beydoun 2007). In light of the overall increase in overweight and obesity in the general
population, it is not surprising that the prevalence of pre-pregnancy overweight and
obesity have also dramatically increased in the 19903 (Kim, Dietz, England, et al. 2007).
Atleasttwostudieshavenotedthat thcprevalenccofobesityhasincreasedatafaster
rate in women of childbearing age (Kim, Dietz, England, et al. 2007; Wang & Beydoun
2007), a phenomenon that may serve to increase the incidence of adverse maternal and
fetal complications.

Both overweight and obesity are more common among individuals of lower
socioeconomic status (SES) and among certain racial/ethnic minorities, particularly
blacks (James, Fowler-Brown, Raghunathan, et al. 2006, Wang & Beydoun 2007; Zhang
& Wang 2004b). As highlighted by the 2003 Institute of Medicine report ‘ Unequal
Treatment: Corgfi'onting Racial and Ethric Disparities in Health Care” (Smedley, Stith,
Nelson, et al. 2003) and other studies (Luo & Waite 2005,. the underlying causes of this
disproportionate burden of overweight and obesity are complex, incompletely understood,

and merit further investigation.

Numerous studies have examined SES factors in childhood and their association
with overweight and obesity in childhood and adulthood (Chu, Rimm, Wang, et al. 1998;
Dietz 1998; Hardy, Wadsworth & Kuh 2000; Jolliffe 2003). Fewer studies have
compared SES factors in childhood and adulthood (Greenlund, Liu Dyer, et al. 1996;
Lahmann, Limner, Gullberg, et al 2000; Giskes, Lenthe, Turrell, et al. 2008; Parson,
Power, Logan, et al. 1999) or evaluated whether changes in SES status from childhood to
adulthood are associated with overweight or obesity (Goldblatt, Moore & Stunkard 1965;
Braddon, Rodgers, Wadsworth, et al. 1986; Blane, Hm't, Smith, et al. 1996; Langenberg,
Hardy, Kuh, et al. 2003; James, Fowler-Brown, Raghunathan, et al. 2006). In the case of
pre-pregnancy BMI, it remains uncertain whether pre-adulthood SES, adult SES, or
changes in SES across the life course are more strongly associated with overweight and
obesity. In addition there are little data on which SES factors are most strongly linked to
pre~pregnancy overweight and obesity. In this study, I sought to address four general
questions: (1) Is there an inverse association between Adult SES and pre-pregnancy
overweight and obesity? (2) Is there an inverse association between pre-Adult SES and
pre-pregnancy overweight and obesity? (3) Is there a positive association between SES
mobility and pre-pregnancy overweight and obesity? (4) Are the above associations

modified by race/ethnicity?

BACKGROUND

0 esi and SES

Most studies have found an inverse association between obesity and SES
(Brunner, Marmot, Nanchahal, et al. 1997; Croft, Strogatz, James, et al. 1992;
Kuskowaka-Wolk & Pergstrom 1993; Jefl'ery, French, Forster, et al. 1991, Leigh, Fries &
Hubert 1992; Sobal & Stunkard 1989; Wamala, Wolk, & Orth-Gomer 1997; Zhang &
Wang 2004a). In an extensive review of studies published from the 19603 through the
mid-19803, Sobal and Stunkard (1989) reported that the majority of studies (75%) on
women from developed countries, not including the US, found an inverse association
between SES and obesity. The majority of studies (93%) on women from the US. also
reported an inverse association between SES and obesity. Interestingly, in a more recent
extensive review of studies fiom 1988.2004, McLaren (2007) also found that the
majority of the studies (63%) on women from highly developed countries, including the
U. S, found a negative association between lower SES and higher body size. It appears
that over the past three decades, the magnitude of the inverse relation between SES and
obesity observed in the US. has declined slightly (Flegal, Harlan & Landis 1988, Zhang
& Wang 2004b; Chang & Lauderdale 2005; McLaren 2007; Wang & Beydoun 2007).
Nonetheless, there is still a significantly higher proportion of individuals of lower SES
compared to those ofhigher SES who are overweight and obese, across all ages (Wang &

Beydoun 2007).

Socioeconomic Mew: A Review

Socioeconomic status is most often operationalized as education, occupation, or
income or as composite of two of the above constructs (McLaren 2007). Many studies
have viewed education, occupation and income as equivalent measures of SES because
they tend to be highly correlated. There is increasing evidence that suggests that
education, occupation and income might have distinct effects on health outcomes, (Sobal
1991; Galobardes, Smith & Lynch 2000; Herd, Goesling, & House, 2007; Kelaher, Paul,
Lambert, et al. 2008), and therefore initial analyses should examine these SES factors
separately. In line with this thinking investigators are increasingly calling for caution
when using these SES indicators interchangeably (Herd, Goesling, & House, 2007; Geyer,
Hemstrom, Peter, at al. 2006). For example, Davey, Hart, Hole, et al. (1998) found that
when education and occupation were modeled together as covariates, occupation was the
most powerful discriminator of all cause mortality. In a more recmt study, Herd,
Goesling and House (2007) found that variations in the onset of firnctional limitations
and chronic health problems were explained more by education than income. On the
other hand, they concluded that the progression of these health problems were more
closely associated with income than education. The studies reviewed below, provide
further insight into the potentially different associations of obesity with education,

occupation, and income.

Education and Obesity
Education is the indicator most often used to define SES (McLaren 2007). Most

studies have consistently found an inverse association between SES and obesity for

women in developed countries (McLaren 2007) such as the US. (F legal, Harlan &
Landis 1988; Leigh, Fries & Hubert 1991; Zhang & Wang 2004b), U.K. (Wardle, Waller
& Jarvis 2002), Sweden (Lahmann, Lissner, Gullberg, et al. 2000) and Finland (Rissanen,
Heliovaara, Knekt, et al. 1991). Some researchers have suggested that education is a
good proxy of adult social class because it is very stable and unlikely to change, whereas
occupation and income might change over a number of years (Zhang & Wang 2004b;
Parson, Power, Logan, et al. 1999). Of course this would not necessarily hold for studies
of young adults.

Studies that have used educational attainment of parents as a measure of SES
have also found an inverse association with obesity in adulthood among offspring. For
example, in a recent study conducted in the US, a low educational level in a woman’s
fatherwasassociatedwithanincreaseriskofherbeingobeseasanadult(Greenlund, Liu,
Dyer, et al. 1996). Interestingly, the women’s mother’s education had no effect.
Altogether, these studies clearly suggest that the educational level of the individual and
perhaps his or her parent impacts a person’s health. Higher educational level seems to
serve as a protective factor against overweight and obesity, while lower educational level

appears to be a risk factor for being overweight or obese.

Occupation and Obesity

Occupation may be the best measure of post-education SES (Krieger, Chen, Coult,
et al. 2005). When occupation has been used to indicate SES, researchers have
consistently found an inverse relation between SES and obesity for women in developed

countries (McLaren 2007). In a study of Swedish women, a lower occupation class in

adulthood was related to higher BMI (Lahmann, Lissner, Gullberg, et al. 2000), and low
occupational class of the women’s parents was also linked to a woman’s risk of obesity,
even after adjusting for her occupational class. In a British cohort study that used a
somewhat simplified categorization scheme for occupational class (manual vs. non-
manual), individuals whose parents were in the manual occupational class were at greater
risk of obesity in adulthood after adjusting for adult SES and educational attainment

(Hardy, Wadsworth & Kuh 2000).

Income and Obesity

Some investigators have proposed that income is a better indicator of SES than
education and occupation, because income more completely captures material resources
and is highly correlated with education and occupation (Zhang & Wang 2004a). In
addition, income directly reflects potential access to healthier living environments such as
improved and safer housing and neighborhoods with increased access to recreational and
physical activity centers and parks as well as grocery stores with a variety of healthier
food selections

A study in the US. reported that mean BMI was inversely associated with family
income for women of child bearing ages 18-34 during the period 1960-1980 (Flegal,
Harlan & Landis 1988). In another study, the authors also found that family annual
income of women ages 18-60 years was inversely associated with risk of obesity in
adulthood (Zhang & Wang 2004a), and data from the 1971-2002 National Health and

Nutrition Examination Surveys (NHANES) showed an inverse relation between income

and BMI throughout the survey period in Black and White women (Chang & Lauderdale

2005).

Other SES Indicators and Obesity

Public assistance status (often times categorized -— yes/ no) and housing situation
(usually categorized — own/rent) are two additional indicators that have been used as
proxies for SES (Wardle, Waller & Jarvis 2002). In the Wardle, Waller & Jarvis (2002)
study, it was found that receiving government financial assistance was associated with
becoming obese while housing situation was not, even after controls (age, marital status,

and ethnicity) were included in the analyses.

SES Mobility and Obesity

Few studies have considered a life course perspective and changes in
socioeconomic status in relation to adult BMI. Parsons, Power, Logan, et al. (1999)
conducted a systematic review on childhood factors associated with adult obesity in
developed countries, including the US. Three studies covered in the review (Goldblatt,
Moore & Stunkard 1965; Braddon, Rodgers, Wadsworth, et al. 1986; Blane, Hart, Smith,
et al. 1996) examined the effects of social mobility over the life course on risk of obesity,
and two of the three studies incorporated women. Overall these studies found the
prevalence of obesity was lower in women experiencing upward mobility as opposed to
downward mobility, while in men no relationship was found between social mobility and

obesity in adulthood.

Two other studies have presented similar evidence of the association between
social mobility and risk of overweight or obesity (Langenberg, Hardy, Kuh, et al. 2003;
James, Fowler-Brown, Raghunathan, et a1. 2006). The results of a population-based
prospective cohort study in Britain indicated tint women whose social class of origin was
low were more likely to be overweight as adults (Langenberg, Hardy, Kuh, et al. 2003).
Among women who experienced upward social mobility, their risk of obesity was
between that of women who remained in the lower social class and that of the women
within the higher social class across the life course. Similarly, the risk of obesity among
women who experienced downward social mobility fell between the risk within the social
class they left, and the risk among the class they joined. Within this social mobility efi‘ect
there was a clear gradient. For example, among women who started in the lowest
occupation level, their reduction in obesity risk was modest if they moved up only to the
next occupation level, and more pronounced if they moved up to the highest occupational
level. This study suggests that the consequences of being in an economically
disadvantaged environment in early childhood can be mitigated if women are provided
opportunities for 11de social mobility. On the other hand, it may be that women who
experience upward mobility are those who are thinner to start with (Gotmaker, Must,
Perrin, et al. 1993), assuming that thinness is socially desirable and results in more
opportunities for promotion and advancement. However, this study did not examine
racial differences.

In a study in Pitt County (North Carolina) by James, Fowler-Brown, Raghunathan,
et al. (2006), the focus was only on African American women. In this study, the risk of

obesity among women who experienced downward or upward mobility over the life-

course did not differ from the risk ofobesity among women whose SES remained high
over the life-course. However, a trend was observed that upwardly mobile women had
the lowest prevalence of obesity compared to women who remained in lower
socioeconomic position over their lives. The authors indicated that the lack of significant
findings may be related to the “small numbers in the referent category ...that prevented
the odds for obesity from reaching conventional levels of statistical significance” (James,
Fowler-Brown, Raghunathan, et al. 2006, p. 559).

In summary, very few studies have used a life course approach to examine the
development of obesity over time and consider if changes in a person’s SES, or
intergenerational social mobility, has an effect on risk of obesity. Even fewer studies
have examined this question separately for racial/ethnic minority populations. Overall,
studies that have examined this research question suggest that upward social class
mobility tends to be inversely associated with risk of obesity, indicating perhaps a

protective effect on health.

Pre- re Overw i i SE

Despite the growing research on the relationship between obesity and SES, a
comprehensive literature review suggested there were no published studies directly
addressing the extent to which pro-pregnancy overweight and obesity are associated with
SES. Previous research is relevant (Sobal & Stunkard 1989; Zhang & Wang 2004a,
McLaren 2007), but pro-pregnancy weight covers a particular window in a woman’s life

and presents unique concerns because the burden of overweight and obesity may directly

10

impact the potential for complications in pregnancy as well as fetal programming for later

effects in ofl‘spring.

Theoretical Framework: The Life Course Perspective

The theoretical framework that guides the proposed study is the life course
perspective. The life course perspective suggests that risks of morbidity and mortality are
affected by the accumulation of biological and social insults throughout a person’s
lifetime (Ben-Shlomo & Kuh, 2002; Halfon & Hochstein, 2002; Krieger 2001; Kuh &
Ben-Shlomo, 1997). Environmental risk factors which are more prevalent among lower-
SES populations such as alcohol abuse and illicit drug use, domestic violence, limited
access to a healthy lifestyle (e. g. recreational physical activity, healthy food), economic
hardships, discrimination, decreased access to health care or to health care of adequate
quality, and other life experiences (i.e., social stressors) acmmulate over time to increase
the risk of pre-pregnancy overweight and obesity. The lack in of economic resources
limits access to health promoting activities and resources such as healthy foods, exercise,
and health care.

The life course perspective is particularly relevant to the experiences of low SES
populations and racial/ethnic minorities, particularly blacks, who to this day continue to
experience more economic and social discrimination that negatively impact the quality of
housing (blacks tend to live in less safe neighborhoods with high crime and
environmental hazards such as the presence of lead and other pollutants), educational
opportunities, and health care accessibility (Williams, 1999). To finther emphasize the

impact of racism and discrimination, Williams indicates “It is generally recognized that

11

there are large racial differences in SES, and health researchers routinely adjust for SES
when examining the race—health association. However, SES is not just a confounder of
racial differences in health but part of the causal pathway by which race affects health.
Race is an antecedent and determinant of SES, and racial differences in SES reflect, in
part, the successful implementation of discriminatory policies premised on the inferiority
of certain racial groups” (Williams, 1999, p. 177). In fact, studies that have adjusted for
SES when exploring racial/ethnic differences in health outcomes have resulted in
potential erroneous conclusions due to measurement problems and confounding
(Kaufman, Cooper & McGee 1997; Morgenstern 1997). Unmeasured biological and
genetic factors have been highlighted as potentially explaining the observed race/ethnic
differences found, thus downplaying the historical discriminatory factors that have played
a strong role in minorities’ current predicament (Mlliams, 1999). Williams (1999)
points out that racial difl‘erences in SES, in particular among Blacks suggest that race is
an antecedent and determinant of SES. Williams & Collins (1995) convincingly argue
that “. . . .major social structures and processes such as racism, acculturation, work,
migration, and childhood SES produce inequalities in health” (Williams & Collins, 1995,

p.349)

R sear h i 11 an H theses

This study’s research question is as follows: To what extent are preaduit SES
(SES of the participant’s parents), adult SES, and SES mobility (changes in SES between
pre-adult and adult stages) associated with pro-pregnancy overweight and obesity among

a cohort of Blacks and White/other women ages 15-47.

12

The study’s main hypotheses are:

. Pro-adult SES will be inversely associated with pro-pregnancy overweight and
obesity.

. Adult SES will be inversely associated with pro-pregnancy overweight and
obesity.

. Women with both low pre-adult and low adult SES will be at greatest risk for
pre—pregnancy overweight and obesity.

. Women who experience upward mobility will be at a reduced risk of pre-
pregnancy overweight and obesity

. Women who experience downward mobility will be at an increased risk of
pre-pregnancy overweight and obesity

. SES will have a less protective effect on risk of pro-pregnancy overweight and
obesity among Blacks when compared to Whites/Other.

13

METHODS

Study Population

This study uses data fiom the Pregnancy Outcomes and Community Health
(POUCH) Study which includes pregmnt women recruited to participate in a study on
biologic and psychosocial factors that affect adverse pregnancy outcomes such as preterm
delivery. Women were recruited from August 1998 to June 2004 fi’om 52 clinics in five
Michigan communities: Flint, Saginaw, Kalamazoo, Lansing and Grand Rapids. Women
were enrolled in gestational weeks 15 through 27 (87 percent before week 25). Inclusion
criteria were singleton pregnancy with no known congenital anomaly, maternal age of 15
or more years, maternal serum alpha-fetoprotein (MSAFP) screen in gestational weeks
15—22, no pre-pregnancy diabetes mellitus, and proficiency in English. Eligible women
were invited to participate at the time of prenatal screening. Of the 3,038 women
enrolled, 19 were lost to follow-up, leaving a cohort of 3,019. Underweight (BMI <18.5
kg/mz) women (n=140) were excluded from analyses. Women who were underweight
were excluded from these analyses for two reasons. The main purpose of the study was
to focus on pro-pregnancy overweight and obesity. In addition, the sample size of the
underweight group was small, 140 women, thereby limiting stable estimates in this group.
Hence, this study used data Earn the remaining 2876 women. These participants were
classified as being of normal weight, overweight, or obese. In all analyses women with
‘normal weight’ serve as the referent group.

The study received approval from institutional review boards at Michigan State

University, Michigan Department of Community Health, and nine community hospitals.

14

Measures

At enrollment participants completed a self-recorded questionnaire and were
interviewed by a trained research assistant (RA). In addition relevant information was
abstracted from prenatal screening databases and screening forms. The data used below

were gathered from these sources.

Dependent Variable - Pro-pregnancy BMI Categories

Pro-pregnancy BMI was calculated by using weight in kilograms divided by
height in meters squared. Pro-pregnancy height and weight were self-reported at study
enrollment. Four BMI categories were created: Underweight (<18.5 kg/mz), normal
weight (18.5-24.9 kg/mz), overweight (25-299 kg/m’) and obese (_>_30 kg/m’). These cut-
points are in keeping with the guidelines developed by the Expert Panel on the
Identification, Evaluation, and Treatment of Overweight and Obesity in Adults convened
by the National Heart, Lung, and Blood Institute’s Obesity Education Initiative in
cooperation with the National Institute of Diabetes and Digestive and Kidney Diseases
(Expert Panel on the Identification, Evaluation, and Treatment of Overweight in Adults,
1998). Underweight women were excluded from analyses. This is discussed in the Study

Population section above.
Independent Variables

The study’s independent variables were ‘Adult SES’, ‘Pre-adult SES’, and ‘SES

Mobility’.

15

Adult SES. This study examined women’s adult SES using current education
level, annual household income, usual ocmpation status, and Medicaid Insurance status.
In addition, information was included on the education level and usual occupational
status of the baby’s father. Education level of the mother and the baby’s father was
categorized as follows: ‘less than high school’, ‘high school’, and ‘greater than high
school’. Mother’s annual household income was assessed by asking how much her total
household income was last year. In this study, this variable was categorized as follows:
‘<$5k to <315k’, ‘315k to <330k’, ‘S30k to <850k’, and ‘SSOK or greater’. Mothers
were also asked if they had Medicaid coverage before or during their pregnancy. This
variable was dichotomized into ‘yes’ and ‘no’ categories.

The variable “usual occupation” was assessed by asking mothers an open-ended
question about their and the baby’s father “usual occupation(s).” Data collected on
mother’s “usual occupation” from the first 1336 participants enrolled were coded by a
professional occupational coder using the US Census Bureau’s 1990 Occupational
Classification System codes. Guided by the initial 1336 coded ‘usual occupations” the
remaining mother’s “usual occupation(s)” were coded and categorized by study research
staff as follows:

0 003-235 professional/managerial/technical = 1

243-389 clerical/sales = 2

403-889 service/blue collar = 3

914 homemaker = 4

0 903-905 and 910, 911 other (military/lretired/student) = 5

909 unemployed = 6

16

o 999 unknown = 7

Categories four through seven of mother’s usual occupation were collapsed due to
small sample size. None of the mothers reported unemployed as their usual occupation,
so that category was eliminated. Thus, the final ‘usual occupation’ categories were: 1=
professional/managerial/technical; 2= clerical/sales; 3= service/blue collar,
4=homemakerlotherlunknown. The classification of the ‘usual occupation’ for baby’ 3
father followed the same steps as that of the mother and resulted in the following six
categories: 1= professional Managerial/technical; 2= clericaVsales; 3= service/blue
collar; 4=homemakerlother”, 5=unemployed and 6=unknown.

Because of the high correlations between the individual Adult SES variables
(range of 0.2 to 0.54), a composite measure of Adult SES, or an Adult SES index, was
created based on 2794 participants. From the original 2876 participants, 82 had missing
data on one or more of the variables needed to create the index. These participants were
excluded from the Adult SES index. In addition, 123 participants had missing data on the
variable ‘father of baby’s education’; however, one also had missing data on the other
SES variables, leaving a total of 122 missing cases that Mer was re-coded as being of
‘low’ education. The rationale for this is twofold: first, maternal lack of knowledge about
the education of the baby's fatha may be suggestive of a potential absentee partner, a
circumstance that may be more characteristic of individuals of lower SES. Second,
preliminary analyses indicated this missing item tracked with a ‘low SES’ within other
comparable SES variables.

Hence, the Adult SES index was created using data from the 2794 participants

with no missing data, and those that were recoded as ‘low’, as described above. The

17

index was created as follows. First, the individual Adult SES indices, (i.e. mother and
father of baby’s current education level and usual occupation, mother’s annual household
income and Medicaid Insurance status) were each dichotomized into 0=10w SES and
1=mid to high SES. The categories that were collapsed to make-up the ‘low SES’
category had similar associations (ORs and 95%CI) with the three-level BMI outcome
variable. The dichotomized Adult SES indices were as follows: mother’s education
level (‘low’ =high school or less, ‘high’ =greater than high school); father of baby’s
education level (‘low’ =high school or less and missing, ‘high’ =greater than high
school); women’s annual household income (‘low’ =<$50K ‘high’ =$50K or greater);
Medicaid coverage before or during mother’s pregnancy (‘low’ =yes, ‘high’=no);
women’s usual occupation (‘low’=clerical/sales/service/ blue collar/homemaker/otha/
unknown, ‘high’ =professional/ managerial/technical); and baby’s father usual
occupation (‘low’ =clerical/sales/service/blue collar/homemaker/other/unemployed/
unknown, ‘high’ =professional/managerial/technical).

The six dichotomized Adult SES indices, mother’s education and occupation;
baby’s father’s education and occupation; Medicaid insured status of the mother and the
annual household income of mother were summed. Adding the six dichotomized adult
SES values resulted in an Adult SES score that ranged from 0 to 6. Low, middle and
high SES categories were constructed fiom the Adult SES index (scores 0-6). This was
done to approximate quartiles as follows: the top 27 .6%tile (score of 0) and lower
29.3%tile (scores 4-6) were categorized as ‘low’ and ‘high’ SES, respectively. The three
middle quartiles (27.7%-70.7%, scores 1-3) were considered the ‘middle’ SES group. In

all analyses the ‘high SES’ is the reference category.

18

One goal was to examine pro-pregnancy BMI in relation to individual SES indices
but interpretations were limited by the high correlation between certain indices; these
correlations ranged from 0.2 to 0.54. The Adults SES composite index addressed
problems of correlation among individual indices and provided a more comprehensive
assessment of SES. There was high-inter item correlation among the six SES indices
composing the Adult SES index (KR-20=0.81).

Pre-adult SES This study included five measures of pre-adult SES: education
level and ‘usual occupation’ of the father and mother of the study participant and family
receipt of public assistance during childhood, (‘yes- no’). Education level was
categorized as ‘less than high school’, ‘high school’, and ‘greater than high school’. The
‘usual occupations’ of participants' parents were coded using the same scheme applied to
the participant and father of the baby (described above). For the father of the participant
the ‘usual occupation’ categories were l=professionallmanagerialltechnical;
2=clericallsales; 3==servicelblue collar, 4=homemaker and 5=other/1memployed/unknown.
For the mother of the participant the ‘usual occupation’ categories were: 1=professioml/
managerial/technical; 2=clericallsales; 3=servicelblue collar, 4=homemaker; 5=other,
6=unemployed and 7=unknown.

Because of the high correlations among the individual Pro-adult SES variables, a
composite measure of Pro-adult SES was also created based on 2773 participants. From
the original 2876 participants, 103 had missing data on one of the variables needed to
create the index. These participants were excluded from the Preadult SES index. In
addition, 119 participants had missing data on the variable ‘mother’s education’ and 450

had missing data on the variable ‘father’s education’. These 569 cases were re-coded as

19

being of ‘low’ education because a respondent who does not know the education of her
mother, or of her fatha, may be suggestive of a potential absentee parent, something that
may be more characteristic of individuals of lower SES. Also, preliminary analyses
showed that the missing items tracked with ‘low SES’ in other comparable SES variables
that had minimal to no missing data.

Hence, the Pre-adults SES index was created using data from the 2773
participants with no missing data, and those that were recoded as ‘low’, as described
above. The process for creating this composite index paralleled that used to create the
Adult SES composite index as described in the previous section. Adding the five
dichotomized pre-adult SES values resulted in a pre-Adult SES score that ranged from O
to 5. The upper 23.3%tile, score of 0 and lower 25.4%tile, scores 3-5 (approximating
quartiles) were categorized as ‘low’ and ‘high’ SES, respectively. The middle 23.4-
74.7%tile, scores 2-3 was considered ‘middle’ SES. Again ‘high SES’ was used as the
reference category. There was high inter—item correlation among the five SES items that
compose the pre-adult SES index (KR-20=0.68)

SES Mobilioz. Mother’s SES mobility was measured by change or lack of change
in SES level from pre-adult to adult, i.e. at the time of enrollment in the study.

Race/eunuch.» Race/ethnicity was assessed by asking each participant to indicate
her racial/ethnic heritage. Individuals were initially categorized into the following
categories: Whites, Blacks, Hispanics, Asians, and Other. For the purpose of this study
the race/ethnicity variable was dichotomized into Black and Whites/Other. The ratiomle

for this categorization is described below.

20

Initial exploratory analyses revealed that the associations between SES and
overweight and obesity were in the same direction and similar in magnitude among
Whites, Hispanics, and those of Other backgrounds. Because of this and because the
sample sizes for Hispanics and Othas were too small to conduct separate analyses, these
individuals were grouped with those of White backgrounds, resulting in a category called
‘Whites/Other’. Further exploratory analyses were conducted comparing the association
between the various SES measures and the dependent variable with and without
Hispanics and those of ‘Other’ backgrounds. The results of these analyses did not differ
suggesting that combining these individuals with Whites in the Whites/Other category
did not affect the results.

Individuals of Asian backgrounds were also added to the ‘White/Other’ category
because the sample size (N=57) was too small to conduct separate analyses. Furthermore,
preliminary analyses revealed that the magnitude of the association between SES and
overweight/obesity among individuals of Asian backgrounds was non-statistically
significant and it was smaller than that observed among the other racial/ethnic groups.
Exploratory analyses were conducted comparing the association between the various SES
measures and overweight and obesity with and without the Asian individuals. As was the
case with the other racial/ethnic variables, the results of these analyses did not difl‘er
suggesting that adding these groups to the Whites/Other category would not affect the
analyses.

Com-iates. Potential confounders included maternal age (in years), parity (first
birth vs. two or more births), and cigarette use in the year prior to pregnancy. In women

the association between age and obesity is bell-shaped with weight increasing up to ages

21

of 55 to 64 years and then decreasing slightly (Wardle, Waller & Jarvis 2002, Lahmann,
Lissner, Gullberg, et al. 2000; Kuskowska—Wolk & Pergstrom 1993). In this study, age
is a continuous variable. High parity has been found to be positively associated with
adult obesity (Lahmann, Lissner, Gullberg, et al. 2000). With every subsequent
pregnancy, weight gain during pregnancy has been a strong determinant in postpartum
weight retention (Kac, Benicio, Velasquez-Melendez, et al. 2004). Parity was assessed
by asking each woman the number of prior live births and for the purpose of this study it
was categorized as a dichotomous variable (0, 21), Finally, cigarette use has been found
to be inversely associated with weight because individuals tend to smoke to control their
appetiteandhencetheirweight, andbecausesmokerswhoquit smokingtendtogain
weight (U .S. Department of Health and Human Services 2004).

Several analyses were conducted to assess if cigarette smoking was associated
with weight. Cigarette use was assessed by asking participants to indicate the number of
cigarettes they usually smoked in a day, week, month or year, within the year before their
pregnancy. One analysis compared the distribution of lifetime smoking status with pre-
pregnancy BMI. A second assessed the association between BMI categories and number
of cigarettes smoked per week in the year before pregnancy. The third compared the odds
of overweight and obesity as a fimction of the number of packs of cigarettes smoked per
day (V: 0, 51/2, >1/2 pack) in the year prior to pregnancy. None ofthese analyses
revealed statistical associations between smoking and preprsgnancy BMI categories in
this cohort. In addition, the smoking variables were added as covariates to models
testing the association between SES and pro-pregnancy weight, but did not confound the

SES-BMI association. Because cigarette use was not a significant confounder explaining

22

the association between SES and the dependent variable in any of the analyses conducted,

this variable was not included in any of the results presented in this paper.

Analfiic Strategy
All analyses were performed using STATA version 9.0. Univariate statistics were

used to describe distributions of each variable. The associations between individual
SES indices, SES composite indices and the dependent variable, pro-pregnancy BMI
categories (normal=refersnt category, overweight, and obese) were evaluated through
multinomial logistic regression models with and without adjustment for age,
race/ethnicity, parity. The analytic sequence consisted of the five steps described below.

First analyses focused on individual adult SES indices (mother and father’s
education level, usual occupation status, annual household income, and Medicaid
Insurance status) in relation to pre-pregnancy BMI categories. The second set of analyses
focused on pre- Adult SES indices (education level and ‘usual occupation’ of the father
and mother of the study participant, and family receipt of public assistance during
childhood) in relation to pre-pregmncy BMI categories and followed the same process
used to assess Adult SES indices.

Inathird step,the SpearmanRanktestwasusedtodetermineiftherewas
multicollinearity among all of the SES indices (results not shown). An examination of the
Spearrnan Rank test suggested thm most of the SES indices were highly correlated. As a
result two separate SES composite indices were created, Adult SES and pro-adult SES

(described in the methods section)

23

The fourth analytic step consisted of using multinomial logistic regression models
to evaluate the associations between the two SES composite indices and the pre-
pregnancy BMI categories. In addition interaction terms were used to empirically test
effect modification by race/ethnicity. In the final fifih step SES mobility was considaed

in association with pre-pregnancy BMI categories.

24

RESULTS
The mean age of the study sample was 26.5 years (Table 1). Approximately 24%
of the women were Black (n=703) and 76% were White/other (1,933 were White, 152
were Hispanic, 47 were Asians, and 41 were Other). Nearly 58% of the women had a

previous live birth.

Socioeconomi statu ov Life Co rse

While 54% of participants had over 12 years of education, only 19.7% had a usual
occupation in the top SES category, i.e. professional/ managerial/ technical (Table 2).
Just under half of the women were insured by Medicaid. A similar percentage reported a
family income of < 30K. Approximately 60% of participants' mothers and 55% of their
fathers completed 12 or fewer years of education and approximately 18% of their
mothers and 18% their fathers had a usual occupation in the top SES category (Table 4).

A history of receiving public assistance as a pre-adult was reported by 37% of study

women.
Pro-pregnancy BMI Categories

The mean pro-pregnancy BMI was 27.1, with 27.4 % classified as obese, 24.5%
as overweight and 48.1% as having normal pro-pregnancy weight (Table 1). A greater
percentage of Black women (34.7%) than White/other women (25%) were obese
(p<0.001). Women with a previous live birth were more likely to be overweight or obese

(p<0.001).

25

Adult SES Indices and new BMI Categories

Pre-pregnancy BMI was significantly associated with mother’s education,
mother’ s usual occupation, Medicaid insurance status, annual household income, and
father of baby’s education and occupation, (p<0.05) when evaluated with a global chi
square test for each of these SES indices (Table 2).

These relations were explored fmther using multinomial regression models (T able
3) and final analyses included maternal age at enrollment, parity (0 vs. 21), and
race/ethnicity (White/other vs. Black) as covariates. Adult SES indices significantly
related to pre-pregnancy overweight (adjusted OR range 1.4-2.1) included mother’s usual
occupation service or bhre collar, father of baby having only high school education and
father’s usual occupation being service, blue collar or unknown (Table 3). Adult SES
indices significantly related to pre-pregnancy obesity (adjusted OR range 1.6-4.8)
included mother’s having only a high school education, mother’s usual occupation being
sales, clerical, service or blue collar, mother’s having received Medicaid insurance before
and during pregnancy, mother’s having an annual income of less than $50 K, and father"

usual occupation being sales, clerical, service, unemployed or unknown.

Pre- SE Indi s an BMI Cat ori
Most of the respondents’ mothers worked in either the savice/blue collar (30%)
or sales/clerical (24%) occupations while most of the fathers worked in service/blue
collar (60%) occupations (see Table 4).
Pre-pregnancy BMI was significantly associated with the participant’s mother’s

usual occupation, history of receiving public assistance in childhood, and father’s

26

education and usual occupation when evaluated by a separate chi square test for each of
the indices (Table 4).

These relations were explored further using multinomial regression models (Table
5) and final analyses included maternal age at enrollment, parity (0 vs. 21), and
race/ethnicity (White/other vs. Black). Pre-adult SES indices significantly related to pre-
pregnancy overweight (adjusted OR range 1.4-1.6) included mother’s father having only
a high school or less than high school education (Table 5). Pre-adult SES indices
significantly related to pro-pregnancy obesity (adjusted OR range 1.3-2.0) included
mother’s mother usual occupation being sales, clericaL service or blue collar, history of
mother receiving public assistance as a child, mother’s father having only a high school
education, less than high school education or education of participant’s father was
missing and mother’s father usual occupation being service, blue collar, other,

unemployed or usual occupation was unknown.

A ult Pre- om sit ES M

Correlations among the adult and pro-adult indices were then assessed using the
Spearman Rank test. Many indices were found to be highly or moderflely correlated
with correlations ranging from 0.2 to 0.54. As described in the Methods section, the SES
indices were dichotomized (T able 6) and summed to create an Adult SES composite

index and a Pro-adult SES composite.

27

Adult and Pre-a_dult Commsite SES Measures and Pre-pregmcy BMI Categories
In examining the adult and pro-adult composite SES and pre-pregnancy BMI,

 

 

conceptual arguments guided the decision to use race/ethnicity-specific models (see
background -theoretical framework: the life course perspective section). Empirical tests
of a race/ethnicity interaction were also explored. Likelihood ratio statistics were used to
compare models with and without the interaction term The improvement of model fit
with the interaction term was not significant for the pre-adult model (p=0.31) but was of
borderline statistical significance for the adult SES model (p=0.07). Also, in the Adult
SES model the race/ethnicity interaction term was significant for pre-pregnancy obesity
(p<0.05). The inverse association between adult SES and risk of pre-pregnancy obesity
was more pronounced in white/other women as compared with that in Black women
(Figure 1).

In analyses that adjusted for maternal age and parity, being in the middle or low
level of the adult SES composite index was significantly associated with an increased risk
of pre-pregnancy obesity in Black women (OR range 2.3-2.8) and in White/other women
(OR range 2.8-3.4) (Tables 7 and 8). For the low and middle levels of the Pro-adult SES
composite index the relation to obesity was more modest among both race/ethnic groups
(OR range 1.5-1.9). This association was statistically significant in Whites/others only,
but this may be due to the larger sample size in this race/ethnic group. The risk of pre-
pregnancy overweight was predominantly linked to being in the low level of the Adult
SES composite index (OR range 1.7-2.0) and again the association reached statistical

significance in the White/other group only.

28

Life Course SES, SES Mobilig and Pre-pmgnangy BMI Categgries
Among Black women, those who moved up to the high Adult SES level did not

have a higher risk of pre-pregnancy overweight or obesity when compared with women
who were always in the high SES level. For example, those who moved fiom low SES in
Pre-Adulthood to high SES in Adulthood were no more likely to be overweight (OR=1.6,
95%CI=0.1-23.8) or obese (OR=0.5, 95%CI=0.0—5.6) than those who were of high SES
in pro-adulthood and adulthood Also, those who moved from mid SES in Pro-Adulthood
to high SES in Adulthood were no more likely to be overweight (OR=2.3, 95%CI=0.4-
14.3) or obese (OR=0.9, 95%CI=0.2-3.3) than those who were of high SES in pre-
adulthood and adulthood (Table 9).Those who were in the high SES group as pro-adults
and moved down to the mid or low level SES as adults were more likely to be overweight
at pro-pregnancy than those who remained in the high SES; the odds ratios for this
association ranged from 4.6-6.8 and varied slightly with the specification of the model,
probably due to small numbers (Table 10).

Among Whites/others the pattern appeared a little different (Tables 11 and 12).
Women who moved from low SES pre-adult to high SES as adult were still at
significantly greater risk for pro-pregnancy obesity (OR=2.9, 95% CI = 1.2-6.8)
compared with their counterparts who were in the high SES group across both the pre-
adult and adult periods. As was true for Black women, the movement of Whites/others
from high SES inpre-adult downto mid orlow SES asadultwas associatedwithan
increased risk of obesity (from high to mid, OR=2.5, 95%CI=0.9-7.0 and fi’om high to
low, OR=2.7; 95%CI=1.7-4.2), though the confidence intervals around some of the

elevated odds ratios included one (Tables 11 and 12). The small samples sizes in these

29

race/ethnicity stratified analyses of SES mobility cannot rule out Type 2 errors due to

limited statistical power.

30

DISCUSSION
This study used a well characterized pregnancy cohort to examine pre-adult and
adult SES indices and their associations with pre-pregnancy overweight and obesity. In
this section, I discuss the pre-adult and adult-SES findings among Black women first and
then women of White/other backgrounds. Then, I discuss the findings of the association
between SES over the life course and pro-pregnancy BMI among Black and White/other

women and conclude with suggestions for fixture research.

Pre- Adu -SES and Pre- BMI Cat ories

Among Black women, pre-adult SES composite index had minimal to no
association with pre-pregnancy overweight or obesity. For obesity the elevated odds ratio
(OR =l.6, 95% CI 0.9-2.8 for mid vs. high and 0R=1.6, 95% CI 0.9-2.9 for low vs. high)
was not statistically significant. This may have been due to limited sample size, but the
odds ratio was similar in magnitude to that of the White/other group. The Greenlund, Liu,
Dyer et al (1996) study found no association between childhood SES (parental education)
and adult obesity in Black women. An additioml study by Stettler, Tershakovec, Zemel
et al (2000) also reported null findings for the relation between childhood SES (maternal
education and household composition) and increased adiposity in Black adult women.
However, these studies contrast with findings fiom the James, Fowler-Brown, Trevillore
et al (2006) and Must, Gortmaker & Dietz (1994) studies, which found that Afi'ican
American women who grew up in the most economically disadvantaged households were
more likely to be obese adults compared to women fi'om less impoverished backgrounds.

The James, Fowler-Brown, Trevillore et al (2006) study defined childhood SES as

31

occupation of family primary earner during childhood and the Must, Gortmaker & Dietz
(1994) study operationalized childhood SES as parental education These inconsistent
finding could be due, in part, to difi‘erences in how SES is measured across studies.

Among Black women, those in the low to middle adult SES groups had a non-
statistically significant increase in risk of prepregnancy overweight (OR=2.0, 95%
CI=O.9-4.7 and OR=1.8, 95% CI=0.8-4.0 respectively) but a significant increase in risk
of pre-pregnancy obesity compared to women in the high SES group. Black women in
the high adult SES group may be engaging in healthier behaviors (e. g. consuming more
fiuits and vegetables and foods with lower fat content, higher levels of physical activity)
than those of women in the lower SES groups (Moreiro & Padrao, 2005). Interestingly,
some prior studies have found that among Black women, adult SES is inversely related to
overweight (Crofi, Strogatz, James et al. 1992) and adult SES has little association with
obesity () (James, Fowler-Brown, Trevillore et al. 2006; Greenlund, Liu, Dyer et al.
1996). Again the different results across studies may in part be explained by variations in
the indices used to measure SES. For example, in the Crofi, Strogatz, James et al (1992)
study, adult SES was measured using a composite of the participant’s education and
occupation. In the James, Fowler-Brown, Trevillore et al (2006) study, adult SES was
measured using a composite of the participant’s education, occupation, employment
status and home ownership. The Greenland, Liu, Dyer et a] (1996) study measured adult
SES by using participant’s education.

Among White/other women in this study, there was no significant association
between pro-adult SES and pie-pregnancy overweight. Women in the low and middle

pre-adult SES groups did experience a non-statistically significant increased risk of pre-

32

pregnancy obesity, relative to high SES women. This is consistent with previous
research; most studies of women fiom European backgrounds have reported an inverse
association between childhood socioeconomic circumstances and body weight in
adulthood (Giskes, Lenthe, Turrell et al 2008; Novak, Ahlgren & Hammarstrom 2006;
Parson, Power, Logan et al 1999). In this study an inverse association also was observed
between adult SES and risk of pre-pregnancy overweight and obesity. Again this is
consistent with other studies of Europeans (Lagenberg, Hardy, Kuh et al 2003; Power,
Manor, Matthews et al 2003; Wardle, Walla & Jarvis 2002; Lawlor, Ebrahim & Smith
2002; Sobal & Stunkard 1989) and of White American women (Ben-Shlomo & Kuh
2002; Greenlund, Lin, Dyer et al 1996) which have shown an inverse relation between
adult SES and obesity. These findings suggest that for certain race/ethnic groups’
socioeconomic circumstances in pie-adulthood and adulthood are associated with
prevalence of obesity in adulthood.

Interestingly, in this study the ‘protective efi‘ect’ of adult high SES appeared
larger among White/other women than among Black women. One explanation might be
that the difi‘erences in lifestyle factors, perceptions of self, and wealth across adult SES
groups and the impact of racial socio-historical context may vary by race/ethnicity. For
example, as suggested by prior research, diet patterns, accessibility to higher quality
foods, and exercising behaviors in high SES blacks may be more similar to those in
lower and mid SES Blacks when compared to the differences in these factors in high
versus low/mid SES Whites/other (Powell, Slater, Mirtcheva et al 2007). Also, Blacks
may pmsemWfihgreateracceptmceofhrga’bodiesmdwithlowermdastmdingofthe
detrimental health consequences of overweight and obesity than White/others

33

(Kumanyika, “filson, Guilford-Davenport 1993), two attitudes that may differentially
impact their risk of pre-pregnancy overweight and obesity. Finally, it is possible that
differences in wealth or accumulated assets (home ownership, stocks) may explain the
differences in the protective efi‘ect of increased SES among the racial groups (Williams,
Collins 1995). That is, White/other women of high SES, or perhaps even at every SES
level, may have significantly more wealth than Black women of high SES, or comparable
SES level. As such, there may be substantial differences in quality of life among Black
and White/other women at every SES level that were not measured or captured in this
study and that may explain the larger protective effect of increased SES among
Whites/other. In addition, the socio-historical context that difl‘ers substantially in the US.

for whites and blacks could impact cultural ideas about nutrition and physical identity.

SES Over the Life Course and flew BMI Categories
Black women who moved down the socioeconomic gradient from pre-adult to

adulthood were more likely to be overweight and obese when compared to those who
maintained high SES over the life course. These findings suggest that downward
mobility may affect women’s pre-pregnancy BMI or that BMI may have some influence
on women’s SES trajectory. There were no significant difi‘erences in risk of pre-
pregnancy overweight and obesity between women who maintained low SES or middle
SES and those who maintained a high SES level over the life course. However, these
null findings may be due to a lack of statistical power as the magnitude of the
associations were large (i.e., adjusted 0R3 ranged fiom 2.5-3.8), but small sample sizes

reduced the power to detect significant differences at conventional levels. These findings

34

are consistent with the findings of the study by James, Fowler-Brown, Raghunathan et al
(2006). Their study detected moderate efl‘ect sizes (i.e., adjusted 0R3 of 16-22) that ,
were not statistically significant.

Among Black women no significant difl‘erences were detected for risk of pre-
pregnancy overweight and obesity among women who improved their SES compared to
those who maintained high SES. In addition, no significant differences were found
among women who improved their SES fiom pro-adult to adult compared to their
counterparts whose SES category remained static. For example, women whose SES over
the life course changed fiom low to high had adjusted odds ratios that indicated a reduced
risk of pre-pregnancy overweight and obesity (i.e., 0.3 and 0.1), essentially a protective
effect, but the upper bound of these confidence intervals included 0R3 of 3.2 and 1.6,
respectively, suggesting that the association could be in eitha direction, increased risk or
protective.

Among Whites/other, experiencing downward mobility over the life course was
associated with an increased risk of pre-pregnancy overweight and obesity when
compared to Whites/other who maintained high SES over their life course. These
findings are consistent with the idea that lowering one’s SES over the life course is
detrimental to one’s BMI. Women with a low pre-adult SES who moved up the SES
ladder to mid or high SES in adulthood also had an increased risk of pre-pregnancy
overweight and obesitywhen comparedtothosewhomaintainedthe most advantaged
SES circumstances throughout the life course. These findings suggest that SES in the pre-
adult period is uniquely associated with pre-pregnancy BMI for some race/ethnic groups.

That is, despite improvements made on socioeconomic circumstances in adulthood,

35

women still do not share the health advantages of those who maintained the most
advantaged SES over their lives.

As was the case with Black women, no significant differences were found among
White/other who improved their SES compared to women who remained in low SES
over their life course. However, White/other women whose SES improved from mid to
high were less likely to be obese than women who remained in mid SES over their lives.
Essentially, this finding suggests that for White/other middle class women, improvements
in SES result in improved BMI when compared to their initial middle SES status, but
does not equal the weight status of those who remained with a high SES over their life

course

Stug Limitations

Several limitations must be kept in mind when interpreting these study findings.
One limitation is that pre-pregnancy weight and height were self-reported and
retrospective, raising concerns about the accuracy of the information. However, the
recall time was relatively short for most women because they were asked about their pre-
pregnancy weight early in pregnancy (1 5-22 weeks gestation) during study emollment.
Another limitation is that the measure of SES mobility over the life-course relied on only
twotimepoints, oneinpre-arhrlthoodandoneinadulthood. IfmoreassessmentsofSES
were available over a person’s lifetime, one could better assess the trajectory of the
association between SES and the weight status over time. In addition, the individual SES
variables utilized to form the Adult and pre-Adult SES composite indices were given
equal weight, suggesting that each SES variable carried equal risk for overweight and

36

obesity. Further research is needed to examine if different SES variables behave
differently in explaining pre-pregnancy BMI and if so to identify appropriate weights to
correspond to their relative influences.

Another limitation is the relatively smaller sample size of Black women. This
limited the ability to interpret observed associations, even though the magnitudes of some
of the odds ratios were similar to those that were statistically significant in the larger,
White/other group. These findings suggest that future research with larger samples of
Black women is needed. Also, the sample sizes for women of other racial/ethnic
minority groups such as Hispanics, Asians, and Others were too small to conduct separate
analyses so these groups were combined with the Whites. In addition, the sample size of
underweight women was too small to conduct separate analyses and appeared
heterogeneous, with an overrepresentation of disadvantaged women and women of Asian
race. An additional limitation of this study is that women with diabetes mellitus were
excluded fiom the parmt study. Diabetes mellitus has been shown to be associated with
obesity (The Diabetes Prevention Program Research Group 2006). Thus the SES and
pre-pregnancy weight findings cannot be generalized to women with pre-pregnancy
diabetes mellitus. Future research is needed to investigate the association between SES
mobility and prepregnancy BMI among these populations. Finally, this study does not
measure eating and physical activity habits and patterns, two sets of behaviors that have
been found to contribute to overweight and obesity.

Notwithstanding these limitations, this is the first study known to the author that
examines the association between SES mobility and pre-pregnancy BMI, a critical

window in a woman’s life because overweight and obesity can negatively impact

37

pregnancy and fetal programming for later effects in offspring, and make weight loss
after pregnancy more challenging. Also, study findings provide additional support for the
importance of SES and the prevalence of overweight and obesity among women of
childbearing age. Unlike most studies of SES and weight which tend to utilize only one
or two indicators of SES (i.e., education and/or occupation), this study utilized a

comprehensive set of indices to measure SES.

Im licati us for Pu lic Health

For an expectant mother, making lifestyle changes to improve her health should
be encouraged at all times, though evidence suggests it ideal if this occurs prior to her
becoming pregnant (Lu 2007; Johnson, Posner, Biermann, et al. 2006; Lu, Kotelchuck,
Culhane, et al. 2006). The current obesity epidemic (Wang & Beydoun 2007) is likely to
contribute to chronic health problems among all women and in particular among
racial/ethnic minority and lower SES women who are highest risk for obesity. It goes
without saying that ongoing public health interventions are needed to continue
encouraging women to adopt or maintain healthy behaviors before and during pregnancy
Mthapardaflarfowsonthosemostaffectedbyhealthdispafifiesthepoorand
minorities. In addition to individual-level public health interventions, however, macro-
social and economic public health interventions are also needed to improve the
socioeconomic and social status of the population in general and of racial/ethnic minority
groups in particular who face the greatest burden of disease. For example, better
employment opportunities can serve as a gateway to increase access to adequate health

insurance and health care, and to overall enhanced quality of life. In addition, improving

38

production and equitable distribution of foods and improving physical environments
making it safer for individuals to engage in physical activities also can enhance quality of

life. In turn, these changes can serve to influence a person’s health.

39

Table 1. Selected demographic characteristies, POUCH study, 1998-2004

 

 

Total Normal Wt Overwt Obese
(N=2876) (n=l ,383) (n=705) (n=788) Sig.
Mean age (sdev) 26.5 (5.8)
Range = 15-47 years
Race, %
Black 24.4 296 (42.1) 163 (23.2) 244 (34.7) “
White/other 75.6 1,087 (50.0) 542 (25.0) 544 (25.0)
Parity, %
0 42.3 658 (54.1) 273 (22.5) 285 (23.4) "
21 57.7 725 (43.7) 431 (26.0) 503 (30.3)
Mean BMI (sdev) 27 .1 (0.8)
Body weight status, %
Obese 27.4
(BMI = 230)
Overweight 24.5
(BMI = 25.0-29.9)
Normal weight 48.1

(BMI = 18.5-24.9)

 

Notes: Differences in percents were compared with the chi-square statistic (‘p<0.05; ”p<0.001).

4o

Table 2. Adult SES indices and pre-pregnancy BMI categories, POUCH study, 1998-2004

 

Total Normal Wt Overwt Obese

 

 

Adult SES Indices (N=2876) (n=1,383) (E705) (F788) Sig
II % ll °/e ll % II ./0

Mother’s (participant) ed

>HS 1,551 54.0 790 50.9 366 23.6 395 25.5 *"

HS 806 28.1 331 41.1 205 25.4 270 33.5

<HS 516 18.0 261 50.6 133 25.8 122 23.7
Mother’s usual occup

prof/mgr/tech 565 19.7 326 57.7 135 23.9 104 18.4 **

sales/clerical 1,095 38. 1 508 46.4 260 23.7 327 29.9

service/blue collar 1,068 37.1 461 43.2 276 25.8 331 31.0

homemaker/other/tmk 148 5.25 88 59.5 34 23.0 26 17.6
Medicaid insured

No 1,513 52.7 795 52.5 368 24.3 350 23.1 “

Yes 1,360 47.3 586 43.1 336 24.7 438 32.2
Mother’s annual hh inc

250 k 895 32.0 508 56.8 221 24.7 166 18.6 **

30 to <50k 539 19.3 239 44.3 139 25.8 161 29.9

15 to <30k 575 20.6 231 40.2 134 23.3 210 36.5

<15 k 789 28.2 357 45.3 196 24.8 236 29.9
Baby’s father education

>l-IS 1,186 41.2 670 56.5 277 23.4 239 20.2 “

HS 1,139 39.6 469 41.2 305 26.8 365 32.1

<HS 428 14.9 191 44.6 96 22.4 141 32.9

Missing 123 4.3 53 43.1 27 22.0 43 35.0
Baby’s father usual occup

prof/mgr/tech 533 18.5 323 60.6 131 24.6 79 14.8 **

sales/clerical 391 13.6 195 49.9 90 23.0 106 27.1

service/blue collar 1,710 59.5 743 43.5 430 25.2 537 31.4

homemaker/other 102 3.6 68 66.7 18 17.7 16 15.7

unemployed 66 2.3 29 43.9 15 22.7 22 33.3

unknown 74 2.6 25 33.8 21 28.4 28 38.8

 

Notes: Differences in percents were compared with the chi-square statistic (‘p<0.05,” p<0.01).
Missing data for the Adult SES Indices ranged from 3 for the variables “Mother’s ed, and
Medicaid uninsured” to 78 for “Mother’s annual household income”. Only the missing cases for
the variable “Baby’s father education” are included in the table because these cases were recoded
as low when the Adult SES Index was constructed. Missing from the other variables are not
includedbecausetheyarenotmodeledinanyofthcanalyses. Thisisdescribedingreaterdetail in
the Adult SES and Pre-Adult SES sub—sections of the “Measures” section of the document.

41

Table 3. Unadjusted and adjusted risk of pie-pregnancy overweight and obesity by adult SES
indices, POUCH study, 1998-2004

 

 

 

 

. Unadjusted ORs Adjusted Ors

Adulth—fggfrces Overwt Obese Overwt Obese

OR 95%CI OR 95%CI OR 95%Cl OR 95%Cl
Mother’s educ
ref >HS
HS 1.3 1.1-1.7 1.6 1.3-2.0 1.3 1.0-1.7 1.6 131.9
<HS 1.1 0.9-1.4 0.9 0.7-1.2 1.1 0.8-1.5 0.9 0.6-1.1
Mother’susualoccup
refprof7mgr/tech
sales/clerical 12 1.0—1.6 2.0 1.6-2.6 1.2 0.9-1.6 2.1 1.6-2.7

service/blue collar 1.4 1.1-1.9 2.3 1.7-2.9 1.4 1.1-1.9 2.3 1.7-3.1
hrnaker/other/rmk 0.9 0.6—1.5 0.9 0.6-1.5 0.9 0.6—1.5 1.0 0.6-1.6
Medicaid insured
ref no
yes 1.2 1.0-1.5 1.7 1.4-2.0 1.2 1.0-1.5 1.7 1.4-2.1
Mother’s annual hh
income
ref 250 k
30 to <50k 1.3 1.0-1.7 2.1 1.6-2.7 1.3 1.0-1.7 2.2 1.7-2.9
15 to <30k 1.3 1.0-1.7 2.8 2.2-3.6 1.3 1.0-1.8 3.0 2.3-4.0
<15 k 1.3 1.0-1.6 2.0 1.6-2.6 1.3 1.0-1.7 2.2 1.6-2.9
Baby’s father educ
ref >HS
HS 1.6 1.3-1.9 2.2 1.8-2.7 1.6 1.3-1.9 2.3 1.8-2.8
<HS 1.2 0.9-1.6 2.1 1.6-2.7 1.3 0.9-1.7 2.4 1.8-3.2
missing 1 .2 08-20 2.3 1.5-3.5 1 .2 0.7-2.0 2.3 1.5-3.6
Baby’s father usual
occup
ref prof/mgr/tech
sales/clerical 1.1 0.8-1.6 2.2 1.6-3.1 1.2 0.8-1 .6 2.4 1.7-3.4
service/blue collar 1.4 1.1-1.8 3.0 2.3-3.9 1.4 1.1-1.8 3.1 2.3-4.1
homemaker/other 0.7 0.4-1.1 1.0 0.5-1.7 0.7 0.4-1.2 1.0 0.6-1.9
unemployed 1 .3 0.7-2.5 3.1 1.7-5.7 l .4 0.7-2.8 3.4 1.8-6.5
tmknown 2.0 1.1-3.7 4.6 2.5-8.3 2.1 1.1-4.0 4.8 2.6-8.9

 

Notes: Bold text is used in the table to more easily identify the ORs that are statistically
significant. The referent eategory is ‘normal weight’. The ‘Adjusted ORs’ colunm represents
results ofanalyses adjusting for maternal age at enrollment, parity (W521), and race/ethnicity
(White/other vs. Black). As described in greater detail in the ‘Adult SES’ sub-section of the
“Measures” section of the document, only the missing cases for the variable “Baby’s father
education”are included inthetablebecausetheseeaseswererecodedas lowwhendreAdult
SES Index was constructed.

42

Table 4. Pie-adult SES indices and pre-pregmncy BMI categories, POUCH study, 1998-2004

 

 

 

Pre- A dult Total Normal Wt Overwt Obese
SES Variable (N=2876) (F1583) JF705) (n=788) Sig
ll ./0 n 70 II '/o n ./0
Mother’s mother ed
>HS 1,038 36.1 535 51.5 239 23.0 264 25.4
HS 1,363 47.4 645 47.3 340 24.9 378 27.7
<HS 356 12.4 150 42.1 96 27.0 110 30.9
Missing 119 4.1 53 44.5 30 25.2 36 30.3
Mother’s mother
usual occup
prof/mgr/tech 523 18.2 287 54.9 124 23.7 112 21.4 "
sales/clerieal 697 24.2 337 48.4 150 21.5 210 30.1
service/blue collar 863 30.0 392 45.4 223 25.8 248 28.7
homemaker 741 25.78 341 46.0 199 26.9 201 27 . 1
other/unempl/tmk 52 1.8 26 50.0 9 17.3 17 32.7
Hx of Public Assist
No 1,745 62.9 887 50.8 440 25.2 418 24.0 "
Yes 1,028 37.1 443 43.1 249 24.2 336 32.7
Mother’s father ed
>HS 838 29.1 466 55.6 181 21.6 191 22.8 *“
HS 1,241 43.1 577 46.5 331 26.7 333 26.8
<HS 347 12.1 141 40.6 90 25.9 116 33.4
Missing 450 15.7 199 44.2 103 22.9 148 32.9
Mother’s father
usual occup
profimgr/tech 503 17.5 287 57.1 116 23.1 100 19.9 "
sales/clerical 241 8.4 131 54.4 46 19.1 64 26.6

service/blue collar 1,724 59.9 789 45.8 446 25.9 489 28.4
hmker/oth/unempl 106 3.7 46 43.4 28 26.7 32 30.2
unknown 302 10.5 130 43.1 69 22.9 103 34.1

 

Notes: Differences in percents were compared with the chi-square statistic (’p<0.05; " p<0.01).
Missing data for the Pre-adult SES Index — “history ofpublic assistance” was 103. Missing fi'om
this variableis notincludedbecause missingdataforthis variablearenotmodeledinanyofthe
analyses. As described in greater detail in the ‘Pre-Adult SES’ sub-section of the “Measures”
section of the document, the eategory ‘Missing’ is included for only the variables “Mother’s
mother education” and “Mother’s father education” because these eases were recoded as ‘low’
education when the Pre-Adult SES Index was constructed.

43

Table 5. Unadjusted and adjusted risk of pre-pregnancy overweight and obesity by pre-adult SES
indices, POUCH study, 1998-2004

 

 

 

 

Unadjusted ORs Adjusted ORs
InPdreee‘siNutzlili) Overwt Obese Overwt Obese
OR 95%CI OR 95%CI OR 95%CI OR 95%CI

Mother’s mother ed

ref >HS

HS 1.2 1.0-1.4 1.2 1.0-1.4 1.2 0.9-1.4 1.2 1.0-1.5

<HS 1.4 1.1-1.9 1.5 1.1-2.0 1.4 1.0-1.9 1.4 1.0-1.8

Missing 1.3 0.8-2.0 1.4 0.9-2.2 1.1 0.7-1.8 1.2 0.8-1.9
Mother’s mother
usual occup

ref prof/mgr/tech

sales/clerical 1.0 0.8-1.4 1.6 1.2-2.1 1.0 0.8-1 .3 1.6 1.2-2.1

srvc/blue collar 1.3 1.0-1.7 1.6 1.2-2.1 1.2 0.9-1.6 1.5 1.1-1.9

homemaker 1.4 1.0-1.8 1.5 1.1-2.0 1.3 1.0-1.7 1.4 1.1-1.9

oth/unempl/unk 0.7 0.3-1.6 1.7 0.9-3.2 0.6 0.3-1.5 1.3 0.7-2.6
Hx of Public Assist

ref No

Yes 1.1 0.9-1.4 1.6 1.3-1.9 1.1 0.9-1.3 1.5 1.2-1.8
Mother’s father educ

ref >HS

HS 1.5 1.2-1.8 1.4 1.1-1.7 1.4 1.2-1.8 1.3 1.1-1.7

<HS 1.6 1.2-2.3 2.0 1.5-2.7 1.6 1.1-2.1 1.9 1.4-2.5

Missing 1.3 1.0-1.8 1.8 1.4-2.4 1.2 0.9-1.7 1.5 1.1-2.1
Mother’s father
usual occup

ref prof/mgr/tech

sales/clerical 0.9 0.6-1.3 1.4 1.0—2.0 0.9 0.6-1.3 1.4 1.0-2.1

srvc/blue collar 1.4 1.1-1.8 1.8 1.4-2.3 1.3 1.0-1.7 1.7 1.3-2.2

other/unempl 1.5 0.9-2.6 2.0 1.3-3.4 1.5 0.9-2.5 1.8 1.1-3.1

rmknown 1.3 0.9-1.9 2.3 1.6-3.2 1.3 0.9-1.8 2.0 1.4-2.9

 

Notes: Boldtext isusedinthetabletomore easily identifytheORsthatarestatistieally
significant. The referent category is ‘normal weight’. The ‘Adjusted ORs’ column represents
results ofanalyses adjusting for maternal age at enrollment, parity (0 vs 21), and race/ethnicity
(White/other vs. Black). As described in greater detail in the ‘Pre-Adult SES’ sub-section of the
“Measures” section of the document, the category ‘Missing’ is included for the variables
“Mother’s mother education” and “Mother’s father education” because these cases were recoded
as ‘low’ education when the Pre—Adult SES Index was constructed.

Tab1e6. Distrrbutionofpre-fiegmncy BMIbycategoriesofhighandlowadult&pe-adultSESindiws,
POUCH study, 1998-2004

 

 

 

Study Normal Wt Overwt Obese
Adult and Pre-Arhlt Population (F1383) (F705) (F788) Si
sas Variables =2876) g
n % a % n % n %
Mother’s ed
>HS H 1,551 54.0 790 50.9 366 23.6 395 25.5 **
SHS L 1,322 46.0 592 44.8 338 25.6 392 29.67
Mother’s usual occup
prof/mgr/tech H 565 19.7 326 57.7 135 23.9 104 18.4 "
sales/c1erica1/svc/ L 2,31 1 80.4 1,057 45.7 570 24.7 684 29.6
blue/lrmker/oth/unk
Medicaid insured
No H 1,513 52.7 795 52.5 368 24.3 350 23.1 "
Yes L 1,360 47.3 586 43.1 336 24.7 438 32.2
Mother’s annual hh S
250 k H 895 32.0 508 56.78 221 24.7 166 18.6 "
<50 k L 1,903 68.0 827 43.5 469 24.7 607 31.9
Baby’s father education
>HS H 1,186 41.2 670 56.5 277 23.4 239 20.2 ”
SES/missing L 1,690 58.8 713 42.2 428 25.3 549 32.5
Baby’s father usual
prof/mgr/tech H 533 18.5 323 60.6 131 24.6 79 14.8 "
sales/clericaUsvc/ L 2,343 81.5 1,060 45.2 574 24.5 709 30.3
blue/hmker/oth/
rmempl/unk
Mother’s mother
education
>HS H 1,038 36.1 535 51.5 239 23.0 264 25.4 *
SHS/missing L 1,838 63.9 848 46.1 466 25.4 524 28.5
Mother’s mother usual
0‘30“?
prof/mgr/tech H 523 18.2 287 54.9 124 23.7 112 21.4 ”
saleslclericallsvc/ L 2,353 81.8 1,096 46.6 581 24.7 676 28.7
bluecollar/hmker/
l/unk
Hx of Public Assist
No H 1,745 62.9 887 50.8 440 25.2 418 24.0 "
Yes L 1,028 37.1 443 43.1 249 24.2 336 32.7
Mother’s father
education
>HS H 838 29.1 466 55.6 181 21.6 191 22.8 ”
SHS/missing L 2,038 70.9 917 45.0 524 25.7 597 29.3
Mother’s father usual
prof/mgr/tech H 503 17.5 287 57.1 116 23.1 100 19.9 "
sales/clerieaUsvc/ L 2,373 82.5 1,096 46.2 589 24.8 688 29.0
bluecollar/hmker/
oth/unemp/unk

 

Notes: H=High; L=Low, Differences in percents were compred with the chi-square statistic
(‘p<0.05; " p<0.01).

45

Table 7. Risk of pre-pregnancy overweight and obesity by prequ SES composite index
and adult SES composite index for Black women (N=675), POUCH study, 1998-2004

 

 

 

 

Unadjusted ORs Adjusted ORs
SES Indices Overwt Obese Overwt Obese
OR 95%CI OR 95%CI OR 95%CI OR 9596C]
Pre-adult
SES
High 1.0 1.0 1.0 1.0
(ref) " " " "
Mid 1.1 0.6-2.0 1.5 0.8-2.7 1.0 0.6-1.9 1.6 0.9-2.8
Low 1.0 0.6-1.9 1.5 0.8-2.6 0.9 0.5-1.7 1.6 0.9-2.9
Adult
SES
High 1.0 1.0 1.0 1.0
(ref) " .. " .-
Mid 1.6 0.7-3.5 1.6 08-3.] 1.8 0.8—4.0 2.3 1.1-4.8
Low 1.8 0.8-3.9 1.6 08-3.] 2.0 0.9-4.7 2.8 1.4-5.9

 

Notes: Bold text is used in the table to more easily identify the ORs that are statistically
significant. The referent category is ‘normal weight’. The ‘Adjusted ORs’ column
represents results of analyses adjusting for maternal age at enrollment and parity

(0 stl).

46

Table 8. Risk of pro-pregnancy overweight and obesity by pre-adult SES composite index
and adult SES composite index for White/other women (N=2124), POUCH study, 1998-

 

 

 

 

2004
Unadjusted ORs Adjusted ORs
SES Indices Overwt Obese Overwt Obese
OR 95%CI OR 95%CI OR 95%CI OR 95%CI
Pre—adult
SES
High 1.0 1.0 1.0 1.0
(ref) " .. .. ..
Mid 1.3 1.0—1.6 1.6 1.2-2.0 1.2 1.0-1.6 1.5 1.2-2.0
Low 1.4 1.0-1.9 1.9 1.4-2.7 1.4 1.0-1.9 1.9 1.4-2.7
Adult
SES
High 1.0 1.0 1.0 1.0
(ref) '— _ .- ..
Mid 1.2 0.9-1.5 2.5 2.0-3.2 1.2 0.9-1.6 2.8 2.2-3.7
Low 1.6 1.2-2.1 2.6 1.9-3.6 1.7 1.2-2.4 3.4 2.4-4.8

 

Notg: Bold text is used in the table to more easily identify the ORs that are statistically
significant. The referent category is ‘nomial weight’. The ‘Adjusted ORs’ column
represents results of analyses adjusting for maternal age at enrollment and parity

(0 vs 21).

47

Table 9. Risk of pre-pregnancy overweight and obesity by SES over the life cause for Black
women (N=642), POUCH study, 1998-2004

 

 

 

 

Unadjusted ORs AdLusted Ors
53.17.33" 0...... 0m 0...... 0m
OR 95%Cl OR 95700 OR 95%Cl 0R 95%Cl
Non-mobile High 1.0 - 1.0 - 1.0 - 1.0 —
(ref) (n=22)
Mobility mid—>high 2.5 0.4-15.3 0.9 0.2-3.3 2.3 0.4-14.3 0.9 0.2-3.3
(n=24)
Non-mobile Mid 3.5 0.7-16.5 1.7 0.6-4.5 3.8 0.8-18.7 2.5 0.9-7.1
(n=136)

Mobility low—rhigh 2.2 0.1-32.5 0.6 0.1-7.1 1.6 0.1-23.8 0.5 0.0-5.6

Magnify? high—valid 4.0 0.8-21.0 1.1 0.3—3.7 4.7 0.9-25.8 1.8 0.5-6.2

Molifififlow—emid 3.8 0.8-19.0 1.7 0.6-5.0 4.1 0.8-21.0 2.6 0.9-7.8

Mogli-ltitliimdalow 4.7 1.0-22.2 2.0 0.7-5.3 5.4 1.1-27.1 3.5 1.2-10.0

Mogfgflgh—olow 5.7 0.9-34.5 0.9 0.2-4.2 6.9 1.1-44.4 1.7 0.4-8.3

NOEIZEZJE? Low 3.5 0.8-16.3 1.6 0.6-4.2 3.9 0.8-19.3 2.8 1.0-7.9
n=

 

Notes: Boldtext isusedinthetabletomoreeesilyidentifythe ORsthatarestatistically
significant. The referent category is ‘normal weight’. The ‘Adjmted ORs’ column represents
results ofanalyses adjusted by maternal age at enrollment and parity (0 vs 21).

48

Table 10. Risk of pro-pregnancy overweight and obesity by SES mobility over the life
course for Black women, POUCH study, 1998-2004

 

 

 

 

SES Mobility Unadjusted ORs Adjusted ORs
over the Overwt Obese Overwt Obese
Lifecourse OR 95%CI OR 95%CI OR 95%CI OR 95%CI
From Low SES
(n=277)
Static Low 1.0 1.0 1.0 1.0

Low to mid 1.1 0.5-2.2 1.1 0.6-2.1 1.0 0.5-2.0 0.9 0.5-1.7

Low to high 0.6 0.1-6.1 0.4 0.0-3.9 0.3 0.0-3.2 0.1 0.0-1.6
From mid SES

(n=314)

Static Mid 1.0 1.0 1.0 1.0

Mid to low 1.4 0.7-2.5 1.2 0.7-2.0 1.4 0.7-2.5 1.4 0.8-2.4

Mid to high 0.7 0.2-2.2 0.5 0.2-1.5 0.7 0.2-2.3 0.3 0.1-1.0
From high SES

(IF-34)

Static High 1.0 1.0 1.0 1.0

High to Mid 4.0 0.8-21.0 1.1 0.3-3.7 4.6 0.7-28.9 1.0 0.2-4.5

High to Low 5.7 0.9-34.5 0.9 0.2-4.2 6.8 0.8-55.1 0.8 0.1-5.2

Notes: Bold text is used in the table to more easily identify the ORs that are statistically

significant. The referent category is ‘normal weight’. The ‘Adjusted ORs’ column
represents results of analyses adjusted by maternal age at enrollment and parity (0 vs 21).

49

Table 11. Risk ofpre-pregnaney overweight and obesity by SES overthe life course for
White/other women (N=2054), POUCH study, 1998-2004

 

 

 

 

Unadjusted ORs Adjusted ORs
85:02:: Overwt Obese Overwt Obese
OR 95%Cl OR 95%Cl OR 95%CI OR 95%Cl
Non—mobileHigh 1.0 - 1.0 - 1.0 -— 1.0 --
(ref)
(n=365)
Mobilitymid—+high 1.1 0.7-1.5 1.2 0.7-1.8 1.0 0.7-1.5 1.1 0.7-1.7
(n=335)
Non-mobile Mid 1.2 0.9-1.7 2.9 2.1-4.2 1.2 0.9-1.7 3.3 2.3-4.8
(n=536)

Mobility low—>high 1.6 0.7-3.6 2.8 1.2-6.6 l .6 0.7-3.6 2.9 1.2-6.8

Mogi-l—ifjiligh—bmid 1.0 0.6-1.5 2.4 1.5-3.6 1.0 0.7-1.6 2.7 1.7-4.2

Mogififjgiw-mtid l .5 0.9-2.3 3.0 1.8-4.8 1.5 0.9-2.4 3.3 2.0-5.3

Moiiifii: aid—flow l .9 1.3-2.8 3.3 2.1-5.2 2.0 1.3-3.1 4.2 2.6-6.8

Mo(bnil=ifyi high—flow 1.3 0.5-3.3 1.8 0.7-4.9 1.5 0.6-3.7 2.5 0.9-7.0

No§?;?i):ge Low 1.5 0.9—2.3 3.1 1.9-5.0 1.6 1.0-2.5 4.1 2.5-6.8
11:

Notes: Bold text is used in the table to more easily identify the ORs thatare statistimlly

significant. The base category for the DV is ‘normal weight’. The ‘Adjmted ORs’ column
represents results ofanalyses adjusted by maternal age at enrollment and parity (0 vs 21).

50

Table 12. Risk of pre—pregnancy overweight and obesity by SES mobility over the life
course for White/other women, POUCH study, 1998-2004

 

 

 

 

SES Mobility Unadjusted ORs Adjusted ORs
over the Overwt Obese Overwt Obese
Lifeeourse OR 95%CI OR 95%Cl OR 95%CI OR 95%CI
From Low SES
(n=369)
Static Low 1.0 1.0 1.0 1.0

Low to mid 1.0 0.6-1.7 1.0 0.6-1.6 0.8 0.5-1.5 0.7 0.4—1.2

Low to high 1.1 0.4-2.6 0.9 0.4-2.2 0.8 0.3-2.2 0.5 0.2-1.4
From mid SES

(n=l,110)

Static Mid 1.0 1.0 1.0 1.0

Mid to low 1.6 1.1-2.3 1.1 0.8-1.7 1.6 1.1-2.4 1.3 0.9-1.9

Mid to high 0.9 0.6-1.2 0.4 0.3-0.6 0.8 0.6-1.2 0.3 0.2-0.5
From high SES

(n=619)

Static High 1.0 1.0 1.0 1.0

High to Mid 1.0 0.6-1.5 2.4 1.5-3.6 0.9 0.6-1.5 2.4 1.5-3.9

High to Low 1.3 0.5-3.3 1.8 0.7-4.9 1.1 0.4-3.0 2.0 0.7-5.9

Notes: Bold text is used in the table to more easily identify the ORs that are statistically

significant. The referent category is ‘normal weight’. The ‘Adjusted ORs’ column
represents results of analyses adjusted by maternal age at enrollment and parity (0 vs 21).

51

Figure 1. Predicted probabilities of obesity reflecting the effect of the interaction of
race/ethnicity with the three-category ‘Adult SES’ variable

Predicted Probabilities - Obesity

 

 

 

r I f

1 1.5 2 2.5 3
Higher number = Increased Current SES Level

 

 

W. Line torWhites/other —+— LineforBladts

 

 

Note. Results of the multinomial logistic regression analysis indicates that the
interaction shown above is statistically significant. The slope is larger for Whites/other
suggesting that as current/adult SES increases, the reduction in the likelihood of
obesity is greater for Whites/other than blacks.

52

REFERENCES

Anderson IL, Waller DK, Canfield MA, et al. Maternal obesity, gestational diabetes, and
central nervous system birth defects. Epidemiol. 2005;16:87-92.

Baeten JM, Bukusi EA, Lambe M. Pregnancy complications and outcomes among
overweight and obese nulliparous women. Am J Public Health. 2001;91:436—440.

Ben-Shlomo Y, Kuh, D. A life course approach to chronic disease epidemiology:
conceptual models, empirical challenges and interdisciplinary perspectives. Int J
Epidemiol. 2002,31 :285-293.

Blane D, Hart CL, Smith GD, et al. Association of cardiovascular disease risk factors
with socioeconomic position during childhood and during adulthood. BM]. 1996;
313:1434-1438.

Bodnar LM, Ness RB, Markovic N, et al. The risk of preeclampsia rises with
increasing pre-pregnaney body mass index. Ann Epidemiol. 2005;15:475-482.

Bodnar LM, Catov JM, Klebanofi‘ MA, et al. Prepregancy body mass index and the
occurrence of severe hypertensive disorders of pregnancy. Epidemiol. 2007 ; l 8(2):234-
239.

Braddon FEM, Rodgers B, Wadsworth MEJ, et al. Onset of obesity in a 36 year birth
cohort study. BMJ. 1986;293:299—303.

Brunner EJ, Mamet MG, Nanchahal K, et al. Social inequality in coronary risk: central
obesity and the metabolic syndrome: evidence from the Whitehall I] study. Diabetologia.
1997;40:1341-1349.

Caban A], Lee D], Fleming LE, et al. Obesity in US workers: The National Health
Interview Survey, 1986 to 2002. Am J Public Health. 2005;95:1614-1622.

Casey BM, Lucas M], Mcintire DD, et al. Pregnancy outcomes in women with
gestational diabetes compared with the general obstetric population. Obstet Gynecol.
1997;90:869-873.

Chang VW, Lauderdale DS. Income disparities in body mass index and obesity in the
United States, 1971-2002. Arch Intern Med. 2005;165:2122-2128.

Crofi JB, Strogatz DS, James SA, et al. Socioeconomic and behavioral correlates of body

mass index in Black adults: The Pitt Country Study. Am J Public Health. 1992;82:821-
826.

53

Dabelea D, Hanson RL, Lindsay RS, et al. Intrauterine exposure to diabetes conveys risks
for type 2 diabetes and obesity: a study of discordant sibships. Diabetes.
2000;49(l2):2208-221 l .

Davey Smith G, Hart C, Hole D, et al. Education and occupational social class: which is
the more important indicator of mortality risk? J Epidemiol Community Health.
1998;52:153-60.

Dietz WH. Health consequences of obesity in youth: Childhood predictors of adult
disease. J Ped. 1998;105:518-25.

Expert Panel on the Identification, Evaluation, and Treatment of Overweight in Adults.
Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and
Obesity in Adults: Executive Summary. Am J Clin Nutr 1998;68:899—917.

Flegal KM, Harlan WR, Landis JR Secular trends in body mass index and skinfold
thickness with socioeconomic factors in young adult women. Am J Clin Nur
1988;48:535-543.

Flegal KM, Carroll MD, Kuczmarski RJ, et al. Overweight and obesity in the United
States: prevalence and trends, 1960-1994. Int J Obes. 1998;22:39-47.

F legal KM, Carroll Ml), Ogden CL, et al. Prevalence and Trends in Obesity Among US
Adults, 1999-2000. JAMA 2002;288:1728-1732.

Galobardes B, Smith GD, Lynch JW. Systematic review of the influence of childhood
socioeconomic circumstances on risk for cardiovascular disease in adulthood. Ann
Epidemiol. 2006;16:91-104.

Geyer S, Hemstrom 0, Peter R et al. Education, income, and occupational class cannot be
used interchangeably in social epidemiology. Empirical evidence against a common
practice. J Epidemiol Community Health. 2006;60:804-810.

Giskes K, Lenthe FJV, Turrell G et al. Socioeconomic position at difl‘erent stages of the
life course and its influence on body weight and weight gain in adulthood: A longitudinal
study with l3-year follow-up. Obesity. 2008 do]: 10. 1038/oby.2008.54.

Goldblatt PB, Moore ME, Stunkard AJ. Social factors in obesity. JAMA.
l9651192(12)21039-1044.

Gortrnaker SL, Must AM, Perrin JM, et al. Social and economic consequences of

overweight in adolescence and young adulthood. N Engl J Med. l993;329(14): 1008-
1012.

54

Greenlund KJ, Liu K, Dyer AR, et al. Body mass index in young adults associations with
parental body size and education in the CARDIA Study. Am J Public Health.
l99686(4):480—485.

Halfon N, Hochstein, M. Life course health development: An integrated framework for
developing health, policy, and research. The Milbank Quarterly. 2002;80:433-479.

Hardy R, Wadsworth M, Kuh D. The influence of childhood weight and socioeconomic
status on change in adult body mass index in a British national birth cohort. Int J Obes.
2000;24:725-34.

Herd P, Goesling B, House J S. Socioeconomic position and health: The differential
effects of education versus income on the onset versus progression of health problems. J
Health Soc Beh. 2007;48:223-238.

James SE, Fowler—Brown A, Raghunathan TE, et al. Life-course socioeconomic position
and obesity in Afiican American women: The Pitt County Study. Am J Public Health.
2006;96(3):554—560.

Jeffery RW, French SA, Forster JL, et al. Socioeconomic status difi‘erences in health
behaviors related to obesity: the Health Worker Project. Int J Obes. 1991 Oct;15(10):689-
696.

Johnson K, Posner SF, Biermann J, et al. Recommendations to improve preconception
health and health care- United States. MMWR Recomm Rep 2006;55(RR—6): 1—23.

Jollifl‘e D. Extent of overweight among US children and adolescents from 1971-2000. Int
J Obes. 2003;28:4-9.

Kac G, Benicio MHDA, Velasquez-Melendez, G et al. Gestational weight gain and
prepregnancy weight influence postpartum weight retention in a cohort of Brazilian
women. J Nutri. 2004;134:661-666.

Kaufman J S, Cooper RS, McGee DL. Socioeconomic status and health in Blacks and
Whites: The problem of residual confounding and the resiliency of race. Epidemiol.
1997;8z621-628.

Kelaher M, Paul S, Lambert, et al. The impact of difl’erent measures of socioeconomic
position on the relationship between ethnicity and health. Ann Epidemiol. 2008;18:351-
356.

Kim SY, Dietz PM, England L, et al. Trends in pre-pregnancy obesity in nine States,
l993-2003. Obesity. 2007 April;15:986-993.

Krieger N, Williams DR, Moss NE. Measuring social class in US public health research:
Concepts, methodologies, and guidelines. Ann Rev Public Health. 1997;18:341-378.

55

Krieger N. Theories for social epidemiology in the 21st century: an ecosocial perspective.
Int J Epidemiol. 2001;30:668-677.

Krieger N, Chen JT, Coult BA, et al. Lifetime socioeconomic position and twins’ health:
An analysis of 308 pairs of United States women twins. PLoS Med. 2005;2(7):e162

Kuh D, & Ben-Shlomo, Y. (Eds). A life course approach to chronic disease
epidemiology. Oxford: Oxford University Press, 1997.

Kumanyika S, Wilson JF, Guilford-Davenport M. Weight-related attitudes and behaviors
of black women. J Am Diet Assoc. 1993;93:416-422.

Kuskowaka-Wolk A, Pergstrom R. Trends in body mass index and prevalence of obesity
in Swedish women 1980-89. J Epidemiol Community Health 1993;47:195-199.

Lahmann PH, Lissner L, Gullberg B et al. Sociodemographic factors associated with long
term weight gain, current body fatness and central adiposity in Swedish women. Int J
Obesity. 2000;24:685-694.

Langenberg C, Hardy R, Kuh D, et al. Central and total obesity in middle aged men and
women in relation to lifetime socioeconomic status: evidence from a national birth cohort.
J Epidemiol Community Health. 2003;57:816-822.

Lawlor DA, Ebrahim S, Davey SG. Socioeconomic position in childhood and adulthood
and insulin resistance: cross-sectional survey using data from British women’s heart and
health study. BMJ. 2002;325:805.

Leigh JP, Fries JF, Hubert I-IB. Gender and race differences in the correlation between
body mass and education in the 1971-1975 NHANES I. J Epidemiol Community Health.
1991;46:191-196.

Lu MC. Recommendations for Preconception. Am Fam Physician. 2007;76:397-400.

Lu MC, Kotelchuck M, Culhane, JF et al. Preconception care between pregnancies: the
content of internatal care. Matern Child Health J. 2006;10(suppl 7): 107-22.

Luo Y, Waite L]. The Impact of Childhood and Adult SES on Physical, Mental, and
Cognitive Well—Being in Later Life. J Geron. 2005;60B(2):S93-S101.

McLaren KD. Socioeconomic status and obesity. Epidemiol Rev. 2007;29:29-48.

Mokdad AH, Ford ES, Bowman BA et al. Prevalence of obesity, diabetes, and obesity-
related health risk factors, 2001. JAMA 2003;289:76-79.

Moreira P, Padrao P. Educational, economic and dietary determinants of obesity in
Portuguese adults: A cross-sectional study. Eating Behaviors. 2005Aug;7:220-228.

56

Morgenstern H. Defining and explaining race effects. Epidemiol. 1997;82609-611.

Must A, Gortrnaker SL, Dietz WH. Risk factors for obesity in young adults: Hispanics,
African Americans and Whites in the transition years, age 16-28 years. Biomed
Pharmacother. 1994;48:143-156.

Novak M, Ahlgren C, Hammarstrom A A life-course approach in explaining social
inequity in obesity among young adult men and women. Int J Obes. 2006;30:191-200.

Parsons TJ, Power C, Logan S, et al. Childhood predictors of adult obesity: a systematic
review. Int J Obes Relat Metab Disord. 1999;232S1-107.

Plagemann A, Harder T, Kohlhofl’ R, et al. Overweight and obesity in infants of mothers
with long-term insulin-dependent diabetes or gestational diabetes. Int J Obes.
1997;21:451-456.

Powell LM, Slater S, Mirtcheva D et al. Food store availability and neighborhood
characteristics in the United States. Prev Med. 2007;44:189-195.

Power C, Manor 0, Matthews S. Child to adult socioeconomic conditions and obesity in
a national cohort. Int J Relat Metab Disord. 2003;27: 1081-1086.

Rissanen AM Heliovaara M, Knekt P et al. Determinants of weight gain and overweight
in adult Finns. Eur J Clin Nutr. 1991;45:419-30.

Samuels-Kalow ME, Funai EF, Buhimschi C et al. Prepregnancy body mass index,
hypertensive disorders of pregnancy, and long-term maternal mortality. Am J Obstet
Gynecol. 2007;197(5):490.el-490.e6.

Smedley, BD., Stith, AY, Nelson, AR, & the Committee on Understanding and
Eliminating Racial and Ethnic Disparities in Health Care. (2003). Unequal treatment
confronting racial and ethnic disparities in health care. Washington, DC: The National
Academy Press.

Sobal J, Stunkard AJ. Socioeconmic status and obesity: a review of the literature. Psychol
Bull. 1989;105:260-275.

Sobal J. Obesity and socioeconomic status: a framework for examining relationships
between physical and social variables. Med Anthropol. 1991;13:231-247

Stettler N, Tershakovec AM, Zemel BS, et al. Early risk factors for increased adiposity: a
cohort study of African American subjects followed from birth to young adulthood. Am J
Clin Nutr. 2000;72:378-383.

The Diabetes Prevention Research Group. Relationship of body size and shape to the
development of diabetes in the diabetes prevention program. Obesity. 2006;14:2107-2117.

57

US. Department of Health and Human Services. (2004). The health consequences of
smoking: A report of the Surgeon General Atlanta, GA: Centers for Disease Control and
Prevention, National Center for Chronic Disease Prevention and Health Promotion,
Office on Smoking and Health. [Available at

http://www.cdcgov/tobacco/sgr/sgr 2004findex.lrt_rr_1]. Accessed on April 6, 2008.

Waller DK, Shaw GM, Rasmussen SA, et al. Prepregnancy obesity as a risk factor for
structural birth defects. Arch Pediatr Adolesc Med. 2007;161:745-750.

Wamala SP, Wolk A, Orth-Gomer K. Determinants of obesity in relation to
socioeconomic status among middle-aged Swedish women. Prev Med. 1997;26:734-744.

Wang Y, Beydoun MA The obesity epidemic in the United States—Gender, age,
socioeconomic, racial/ethnic, and geographic characteristics: A systematic review and
meta-regression analysis. Epidemiol Rev. 2007;29:6—28.

Wardle J, Waller J, Jarvis MJ. Sex differences in the association of socioeconomic status
with obesity. Am J Public Health. 2002 Aug;92:1299-1304.

Williams DR. Race, socioeconomic status, and health: the added effects of racism and
discrimination. Ann N Y Acad Sci. 1999;896:173-188.

Williams DR, Collins C. US Socioeconomic and racial differences in health: Patterns and
explanations. Annu Rev Sociol. 1995;21:349-386.

Zhang Q, Wang Y. Socioeconomic inequality of obesity in the United States: do gender,
age, and ethnicity matter? Soc Sci Med. 2004a;58: 1 171-80.

Zhang Q, Wang Y. Trends in the association between obesity and socioeconomic status
in US. adults: 1971 to 2000. Obes Res. 2004b;1211622-32.

58

umgrmrgrrrgrrrrrrrt