Title: Crossnational comparison of the association between race and mean waist circumference in older women
1Cross-national comparison of the association
between race and mean waist circumference in
older women
- Kiarri N. Kershaw, Ana V. Diez Roux, Sarah
Burgard - University of Michigan, Ann Arbor
- Department of Epidemiology
- World Congress on Public Health
- April 30, 2009
2Study goal
- To understand the interplay of race and
socioeconomic inequality in Sao Paulo, Brazil and
Havana, Cuba compared with the US and their
associations with abdominal obesity
3Obesity Trends Among U.S. AdultsBRFSS, 1990,
1998, 2007
(BMI ?30, or about 30 lbs. overweight for 54
person)
1998
1990
2007
Source CDC Behavioral Risk Factor Surveillance
System
4Obesity in Developing Countries
Popkin 2006
5Socioeconomic status and obesity
- Understanding the social patterning of obesity is
important for the development of effective
intervention strategies - In high income countries there is typically a
clear inverse association between SES and
obesity1 - In developing countries this relationship is less
clear1,2
1McClaren 2007 2Monteiro 2001
6Race and Obesity
Source Okosun 1999
7Race, SES, and Obesity
- In the US
- Socioeconomic inequality is patterned by race1
- Race and low SES are associated with obesity2
- In Cuba and Brazil
- Arguably less racial inequality than in the US3,4
- Socioeconomic patterning of obesity may be
different
1Williams 1995 2Wang 2007 3Telles 2004 4Fuente
2005
8Specific research questions
- Is there less socioeconomic inequality
(education) by race in Sao Paulo, Brazil and
Havana, Cuba compared with the US? - Is the association between race and mean waist
circumference smaller in Sao Paulo and Havana
than in the US? - Is the association between education and mean
waist circumference different across the
countries?
9Sample
- Women aged 60 years and older
- 1999-2000 Survey on Health and Well-being in
Latin America and the Caribbean (SABE) - 1999-2002 National Health and Nutrition
Examination Surveys (NHANES)
10Study measures
- Waist circumference measured in cm
- Self-identified race
- Age modeled in 5-year age categories
- Education country-specific tertiles
11Methods
- Descriptives
- Multinomial logistic regression (for education
outcome) and linear regression (for waist
circumference outcome) - Interaction terms used to assess significant
country variation in the associations between
race and mean waist circumference and
socioeconomic status and mean waist circumference
12Specific research questions
- Is there less socioeconomic inequality
(education) by race in Sao Paulo, Brazil and
Havana, Cuba compared with the US? - Is the association between race and mean waist
circumference smaller in Sao Paulo and Havana
than in the US? - Is the association between education and mean
waist circumference different across the
countries?
13Logit(Educ) b0 b1Age b2Black b3Brazil
b4Cuba b5Black_Brazil b6Black_Cuba
plt0.05 plt0.01 plt0.10 for interaction term
plt0.05 for interaction term
14Specific research questions
- Is there less socioeconomic inequality
(education) by race in Sao Paulo, Brazil and
Havana, Cuba compared with the US? - Is the association between race and mean waist
circumference smaller in Sao Paulo and Havana
than in the US? - Is the association between education and mean
waist circumference different across the
countries?
15WC b0 b1Age b2Black b3Brazil b4Cuba
b5Black_Brazil b6Black_Cuba
plt0.05 plt0.01 plt0.10 for interaction term
plt0.05 for interaction term
16Specific research questions
- Is there less socioeconomic inequality
(education) by race in Sao Paulo, Brazil and
Havana, Cuba compared with the US? - Is the association between race and mean waist
circumference smaller in Sao Paulo and Havana
than in the US? - Is the association between education and mean
waist circumference different across the
countries?
17WC b0 b1Age b2Black b3Brazil b4Cuba
b5Low b6Med b7Brazil_Low b8Brazil_Med
b9Cuba_Low b10Cuba_Med
plt0.05 plt0.01 plt0.10 for interaction term
plt0.05 for interaction term Statistically
significant trends in Havana and the US
18Conclusions
- Findings
- There was evidence of socioeconomic inequality
patterned by race in Sao Paulo and the US - There was no Black-White disparity in waist
circumference in Sao Paulo, Brazil but there was
in the US - There was no Black-White disparity in waist
circumference in Havana and the socioeconomic
patterning was in the opposite direction compared
with in the US - Potential implications
- Reaffirms that socioeconomic inequality plays an
important role in shaping racial health
disparities in the US - The socioeconomic patterning of obesity in Havana
suggests that the target groups for future
prevention and intervention efforts may be
different from those in the US
19Thank you!
- Co-authors
- Ana Diez Roux
- Sarah Burgard
- Colleagues
- Amar Hamoudi
- Sandra Albrecht
- Brady West
- Funding
- Rackham Merit Fellowship
- Rackham Travel Award
20Extra slides
21Race in Brazil
- Race refers to skin color or physical appearance
more than ancestry - Approximately 53.7 White 38.5 Mulatto 6.2
Black1 - Race relations appear to be better than in the
US2 - Debate in the literature over the extent of
racial inclusivity vs. exclusivity in Brazil2
1CIA World Factbook 2Telles 2004
22Race in Cuba
- Race refers to skin color and physical appearance
more than ancestry - Approximately 65.1 White 24.8 Mulatto or
Mestizo 10.1 Black1 - Argued that there is less racial inequality than
in the US2
1CIA World Factbook 2Fuente 2005
23Exclusions from Brazil sample
- 765 male and 1089 female black and white SABE
Brazil participants - 631 men (82.5) and 920 women (84.5) were used
in the waist circumference analyses - 631 men (82.5) and 916 women (84.1) were used
in the obesity analyses - A different number of participants were excluded
from each of the socioeconomic indicator-specific
analyses based on whether or not they were
missing that particular measure - 6 men and 5 women were excluded from the
education analyses - 52 men and 43 women were excluded from the income
analyses - 3 women were excluded from the occupation
analyses.
24Exclusions from Cuba sample
- 702 men and 1190 black and white participants
- 629 men (89.6) and 1046 women (87.9) were
included in the waist circumference analyses - 625 men (89.0) and 1037 women (87.1) were
included in the obesity analyses - A different number of participants were excluded
from each of the socioeconomic indicator-specific
analyses based on whether or not they were
missing that particular measure - 3 women were excluded from the education analyses
- 5 men and 7 women from the income analyses
- 10 men and 11 women from the occupation analyses.
25Exclusions from US sample
- 1316 male and 1412 female black and white
participants 60 - 1087 men (82.6) and 1098 women (77.8) were
included in the waist circumference analyses - 1084 men (82.4) and 1096 women (77.6) were
included in the obesity analyses - A different number of participants were excluded
from each of the socioeconomic indicator-specific
analyses based on whether or not they were
missing that particular measure - 4 men and 3 women were excluded from the
education analyses - 111 men and 140 women were excluded from the
income analyses - 63 men and 54 women were excluded from the
occupation analyses.
26(No Transcript)