The Gift That Keeps on Giving:

1
The Gift That Keeps on Giving Racial
Inequalities, Socioeconomic Status Health
Across the Lifecourse Cynthia G. Colen, PhD,
MPH September 22, 2008
2
Life Expectancy At Birth, By Race Gender
U.S. Census Bureau. 2007. Statistical Abstract of
the United States 2007.
3
Infant Deaths Per 1,000 Live Births in U.S.
National Center for Health Statistics. 2007.
Health, United States, 2007.
4
Age Adjusted All-Cause Mortality Rates Per 1,000
(18) in U.S., 2005
Source National Vital Statistics System (NVSS),
2005.
5
Age Specific All-Cause Mortality Rates Per 1,000
(18-64 y.o.) in U.S., 2005
Source National Vital Statistics System (NVSS),
2005.
6
Percent of U.S. Adults (18) Who Rate Health As
Fair/Poor, Adjusted for Age, 2004-2006
Source National Health Interview Survey (NHIS),
2004-2006.
7
Percent of U.S. Adults (18) Who Rate Health As
Fair/Poor, 2004-2006
Source National Health Interview Survey (NHIS),
2004-2006.
8
All-Cause Mortality per 100,000 Among Children in
U.S.
Source National Vital Statistics System (NVSS),
2003-2005.
9
Percent of Women in U.S. with Hypertension
Source National Health Nutrition Examination
Survey (NHANES IV), 2003-2006.
10
Probability of Hypertension Among U.S. Women,
1999-2002
Geronimus et al. In Press. Black-White
Differences in Age Trajectories of
Hypertension Prevalence Among Adult Women Men,
1999-2002. Ethnicity and Disease.
11
Complicated Relationship Between Race SES

• Racial disparities in health often more
  pronounced at higher levels of SES
• True for variety of outcomes across lifecourse
  (esp. in infancy mid-life)
• Why do we see this pattern? Can lifecourse
  perspective help us understand unexpected
  relationship?

12
Percent U.S. Adults (18) Who Rate Health As
Fair/Poor
Source National Health Interview Survey (NHIS),
2004-2006.
13
Percent U.S. Infants Weighing lt 2500g at Birth,
2002
US DHHS, CDC, NCHS, Division of Vital Statistics,
Natality Public Use Data 1995-2002, on CDC WONDER
On-line Database.
14
Two Possible Explanations

• Middle-class African Americans more likely to be
  poor in childhood
• OR
• Middle-class African Americans gain fewer health
  benefits from lifetime gains in SES (ie. upward
  mobility)

15
Two Possible Explanations

• Exposure to poverty in childhood negatively
  impacts adult health (cumulative lifecourse
  model)
• Diminishing health returns to upward mobility for
  African Americans (interactive lifecourse model)
• Problem current lifecourse models tend to
  ignore interactions!

16
Existing Lifecourse Models

• Accumulation of Risk (Cumulative Class) Model
• Chains of Risk (Pathways) Model
• Critical Periods (Latency) Model

17
Accumulation of Risk Model

• SES typically captured using occupation (ie.
  civil servants in Great Britain)
• Measured during childhood, at labor market entry,
  in adulthood
• Longer you are exposed to low SES, worse off your
  physical health is

18
Chains of Risk Model

• Focus is on transitions from one place in
  socioeconomic hierarchy to another
• If individual born into poor family, most likely
  to adhere to certain trajectory
• Least tested lifecourse model

19
Critical Periods Model

• Certain periods during lifecourse that impact
  long-term health outcomes
• To date, focus has been on intrauterine
  environment
• Restriction of resources (maternal nutrition)
  leads to growth retardation ultimately poor
  adult health
• Empirical evidence is mixed - methodological
  problems

20
What Existing Lifecourse Models Tell Us

• 2 of 3 models assume linear, cumulative
  association between SES health
• Upward mobility will result in better health
  downward mobility will lead to worse health
• Fetal origins hypothesis allows for more
  complexity but focus is on intrauterine
  environment maternal nutrition

21
Conceptual Shortcomings of Existing Lifecourse
Models

• Upward mobility may have restricted health
  returns for minorities
• Largely due to structural individual racism
• Challenges linear, cumulative relationship
  between SES health

22
Methodological Shortcomings of Existing
Lifecourse Models

• Majority of studies rely solely on data from
  Western Europe
• Whether lifetime SES health varies by race
  rarely assessed
• Ignore how gender may influence impact of SES on
  health over time
• Maternal and infant wellbeing not typically
  endpoints of interest

23
How Should Existing Lifecourse Models Be
Strengthened?

• Consider how
• Childhood exposures interact with socioeconomic
  conditions in adulthood
• Dynamic processes shape health over time
• Contextual factors influence probability that
  risk of disease or death decreases as SES
  increases

24
What Is (Desperately) Needed?

• Create lifecourse models that reflect realities
  of minority women
• Consider how inequalities interact across race,
  class, gender to influence health
• Do not forget women are often key players in
  social networks

25

• Structural Inequalities
• Restricted Returns to Education
• Labor Market Segmentation
• Residential Segregation
• Differential Wealth Accumulation

• Psychosocial Stressors
• Interpersonal Discrimination
• Unique Coping Styles
• (ie. John Henryism)
• Dual/Competing Identities

Lifetime SES
Physical Health Status

• Weathering Hypothesis
• Physical deterioration
• Delayed fertility timing
• Kin network composition

26

• Structural Inequalities
• Restricted Returns to Education
• Labor Market Segmentation
• Residential Segregation
• Differential Wealth Accumulation

• Psychosocial Stressors
• Interpersonal Discrimination
• Unique Coping Styles
• (ie. John Henryism)
• Dual/Competing Identities

Lifetime SES
Maternal Infant Health

• Weathering Hypothesis
• Physical deterioration
• Delayed fertility timing
• Kin network composition

27

• Structural Inequalities
• Restricted Returns to Education
• Labor Market Segmentation
• Residential Segregation
• Differential Wealth Accumulation

• Psychosocial Stressors
• Interpersonal Discrimination
• Unique Coping Styles
• (ie. John Henryism)
• Dual/Competing Identities

Lifetime SES
Maternal Infant Health

• Weathering Hypothesis
• Physical deterioration
• Delayed fertility timing
• Kin network composition

28
Is This What I Am Proposing?
Whites
Lifetime SES
Blacks
Health
29
Kind ofbut not exactly

• Relationship between SES health over lifecourse
  likely to be different across racial/ethnic
  groups
• Middle-class status provides less beneficial
  rewards for certain racial/ethnic groups
• Upward mobility likely to be qualitatively
  different experience for African Americans
  compared to Whites. But why?

30

• Structural Inequalities
• Restricted Returns to Education
• Labor Market Segmentation
• Residential Segregation
• Differential Wealth Accumulation

Lifetime SES
Maternal Infant Health
31
Racial Residential Segregation

• Blacks substantially less likely than Whites with
  similar sociodemographic characteristics to move
  from poor to nonpoor census tracts (Crowder
  South 2005)
• Live in communities that serve as buffer
  between poor Black areas White areas
  (Pattillo-McCoy 1999)
• African American women living in census tracts
  with highest concentration of Black residents are
  2.7 times more likely to die (Jackson et al. 2000)

32
Net Worth by Income and Race, 1999
Shapiro. 2005. The Hidden Costs of Being African
American.
33
Net Worth by Middle Class Status Race, 1999
Shapiro. 2005. The Hidden Costs of Being African
American.
34

• Psychosocial Stressors
• Interpersonal Discrimination
• Unique Coping Styles
• (ie. John Henryism)
• Dual/Competing Identities

Lifetime SES
Maternal Infant Health
35
Percent Reporting Racial Discrimination, By Race
(NSFH)
Ren et al. 1999. Racial/Ethnic Disparities in
Health The Interplay Between Discrimination and
SES. Ethnicity Disease 9151-165.
36
Competing Identities

• Upwardly mobile African American women may need
  to adopt dual often competing identities (Bell
  Nkomo 2001 Jones Shorter-Gooden 2003)
• To navigate cultural social distance between
  families/communities of origin and achieved SES
• Maintaining dual identities could lead to
  conflict stress when women forced to suppress
  key aspects of one of these identities

37
Lifetime SES
Maternal Infant Health

• Weathering Hypothesis
• Physical deterioration
• Delayed fertility timing
• Kin network composition

38
Weathering Hypothesis

• The health of African American women begins to
• decline in early adulthood
• deteriorates at an accelerated rate
• as a physical manifestation of social, economic,
  and political exclusion

39
Probability of Hypertension Among U.S. Women,
1999-2002
Geronimus et al. In Press. Black-White
Differences in Age Trajectories of
Hypertension Prevalence Among Adult Women Men,
1999-2002. Ethnicity and Disease.
40
Timing of Reproduction
Exponential Increases In Maternal
Infant Morbidity Mortality
Delayed Childbearing
Accelerating Aging (Weathering)


41
Putting My Money Where My Mouth Is

• Enough theorizing already
• Lets test the idea that upward mobility will
  have restricted MCH health returns for African
  Americans
• We want some empirical proof!

42
Research Hypotheses

• Among Whites, probability of LBW will be lower
  among upwardly mobile women compared to
  chronically poor counterparts
• Among Blacks, probability of LBW will not be
  lower for upwardly mobile women compared to
  chronically poor counterparts
• Black-White differences in association between
  upward mobility LBW will not be explained by
  proximate maternal behavioral risk factors

43
Three Data Sources

• National Longitudinal Survey of Youth 1979
  (NLSY79)
• National Longitudinal Survey of Youth 1979
  Childrens Supplement
• 1970 Public Use Microdata Samples from the U.S.
  Decennial Census (PUMS)

44
Description of NLSY79

• NLSY79 includes data from 12,686 young men
  women who were 14-22 in 1979
• Only longitudinal U.S. dataset to include
  multigenerational measures of SES maternal
  health behaviors
• Retention rates range from 96 in 1983 to 77.5
  in 2002

45
Inclusion Criteria

• NonHispanic Black and White women who were 14-22
  in 1979
• Had at least one child by 2002
• Lived in a household at age 14 in which
  income-to-needs ratio did not exceed 200 of
  poverty

46
Upwardly Mobile
Lifetime Income
Chronically Poor
T1
T2
Time
47
Data Includes Multiple Generations

• 1st Generation Adult male/female in household of
  NLSY respondent when they were 14 years of age
• Grandparents!
• 2nd Generation NLSY respondent who gave birth to
  at least one child before 2002
• Moms!
• 3rd Generation Children for whom birthweight was
  reliably obtained
• Kiddos!

48
Dependent Variable

• Low birthweight
• 1 lt 2500 grams
• 0 2500 grams
• Obtained through maternal recall rather than
  self-report
• No difference in proportion of missing
  birthweight data across racial categories (8.5
  for Whites 8.7 for Blacks)

49
Independent Variables

• Grandparents SES combined education
  occupation data from NLSY79 with income data from
  1970 Public Use Microdata Samples (PUMS) to
  predict income-to-needs ratio
• Maternal SES continuous measure of household
  income during the year mothers became pregnant
• All household income adjusted for inflation and
  reported in 2002 dollars

50
Analytic Strategy

• Series of logistic regression models estimated
  separately for Blacks and Whites
• LnPr(y 1x) / 1-Pr(y 1x) ß0 ß1x1
  ß2x2 ßnxn e
• Robust SEs adjusted for clustering at level of
  original NLSY79 respondent

51
(No Transcript)
52
(No Transcript)
53
(No Transcript)
54
(No Transcript)
55
(No Transcript)
56
(No Transcript)
57
Significant Predictors of LBW for Whites Who Were
Poor in Childhood
58
Significant Predictors of LBW for Blacks Who Were
Poor in Childhood
59
Conclusions

• Among Whites, intergenerational gains in SES
  associated with almost 50 reduction in odds of
  LBW
• Among Blacks, upward mobility does not appear to
  curtail the likelihood of LBW
• Muted effect of adult SES for Blacks could not be
  explained by maternal health behaviors

60
Conclusions

• For African Americans, social networks
  especially those maintained through familial ties
  are important health promoting mechanisms
• Among Blacks, having a coresidential grandmother
  was associated with a 53 reduction in the odds
  of LBW
• This is net of effects of family income!

61
Maintaining Kin Networks
Upward Mobility
62
What Does This Mean for MCH Practice?

• It is obvious that rising tide will not lift all
  boats equally
• Expanding focus of certain programs to include
  nonpoor women
• Cannot just focus on proximate determinants
  (maternal health behaviors) - have to consider
  structural factors as well

63
What Does This Mean for MCH Practice?

• Need to consider unintended consequences of
  programs policies
• If we ask minority women - chronically poor or
  upwardly mobile - to delay childbearing, will MIH
  outcomes be worse?
• If we ask them to delay childbearing, will key
  members of social networks be able (healthy
  enough) to help with childrearing?

64

• Thank You!