Transcript Slide 1

Does the County Poverty Rate Influence Birth Weight and Infant Mortality in Kansas?
Kevin Kovach, DrPH(c), MSc, CHES
Johnson County Department of Health and Environment – Olathe, Kansas
Introduction
Social economic status (SES) is a known risk factor for
almost all health conditions and is considered a fundamental
cause of health. The fundamental cause theory suggests that
SES embodies a range of resources, such as money,
knowledge, and power, and that individuals living with low
SES lack the resources to improve their health. According to
the ecological framework, SES can act both at the individual
level (e.g., income or education) and at the community level
(e.g. the built environment, social systems, or culture) [1].
The infant mortality rate is a sensitive measure that not only
measures the health status of infants, but is a leading
indicator of broader public health issues [2]. Birth weight is a
primary risk factor for infant mortality and has also been
proposed as a public health indicator [3]. The purpose of this
study was to examine if community level SES,
operationalized as the county poverty rate, has an
independent effect on infant mortality and birth weight, after
taking into account individual maternal SES and behaviors.
This was accomplished through statistical analysis of linked
birth and death certificate data from 2006 to 2010.
Methods
Data
Conclusion
The county poverty rate was observed to have an indirect, but not a direct effect and infant mortality and birth
weight. Risk factors with the most direct influence include: 1) race, 2) marital status, 3) cigarette smoking, and 4)
alcohol use. Both education and smoking status appeared to interact with the county poverty rate.
Table 1: Effect of County Poverty Rate on Birth Weight and Infant Mortality
Birth Weight
Independent Variables
Infant Mortality
Crude
β (95% CI)
Adjusted
β (95% CI)
Crude
OR (95% CI)
Adjusted
OR (95% CI)
County Poverty Rate
-6 (-7; -6)
-2 (-2; -1)
1.02 (1.01; 1.03)
1.00 (0.99; 1.01)
Age
10 (10; 11)
5 (4; 6)
0.96 (0.95; 0.97)
0.99 (0.98; 1.01)
Race
• White
• Black
-234 (-243; -224)
-181 (-192; -172)
2.02 (1.68; 2.43)
1.50 (1.24; 1.83)
Education
• B.S. or greater
• A.S. or some college
• H.S. or GED
• Less than H.S.
-80 (-86; -74)
-153 (-160; -147)
-177 (-184; -169)
-16 (-23; -9)
-47 (-55; -39)
-63 (-73; -54)
1.44 (1.19; 1.73)
2.19 (1.82; 2.63)
2.56 (2.13; 3.08)
1.18 (0.85; 1.20)
0.99 (0.77; 1.28)
1.20 (0.99; 1.44)
Health Insurance
• Private
• Medicaid
• Self
-145 (-151; -139)
-80 (-88; -73)
-12 (-19; -5)
25 (14; 35)
1.76 (1.53; 2.02)
1.57 (1.26; 1.95)
1.01 (0.85; 1.20)
0.99 (0.77; 1.28)
Married (no)
-147 (-152; -142)
-41 (-47; -34)
1.88 (1.67; 2.13)
1.26 (1.08; 1.46)
Cigarette Smoking (yes)
-177 (-183; -171)
-132 (-139; -126)
1.74 (1.53; 1.99)
1.36 (1.18; 1.58)
Alcohol Use (yes)
-233 (-276; -171)
-112 (-163; -60)
2.57 (1.15; 5.77)
1.55 (0.68; 3.50)
Figure 1: Birth Weight Predicted by the County Poverty Rate
The county poverty rate was observed to have a small but
consistent negative effect on birth weight and infant mortality
in Kansas. After adjusting for maternal characteristics the
effect of the county poverty rate was attenuated toward the
null. Maternal characteristics that remained significantly
associated with birth outcomes after adjustment included: 1)
race, 2) education, 3) marital status, and 4) smoking status.
Using the social epidemiological principal of “risk regulators,”
it could be proposed that the county poverty rate acts as a
proxy to some social phenomenon that drives more proximal
risk factors (e.g., less education, out of wedlock pregnancy,
smoking, alcohol use, etc.). In this case, addressing the social
context in these neighborhoods may be more beneficial than
addressing individual level risk factors.
Future studies could be improved by using smaller
geographical units of measurement, such as census tracts.
Qualitative methods could be used to better define the social
context. Community based participatory research may a good
method for addressing this issue in an action oriented
manner.
Sources
Figure 2: Probability of Infant Death Predicted by the County Poverty Rate
This was a cross sectional, multilevel analysis of birth and
death certificate data (2006 – 2010) linked with data from the
American Community Survey 2006 – 2010 five year
estimates. Outcomes included birth weight and infant
mortality. Risk factors included the county poverty rate and
maternal characteristics (Table 1).
1. Glass, T., & McAtee, M. (2005). Behavioral Science at the Crossroads in Public Health:
Extending Horizons, Envisioning the Future. Social Science and Medicine, 1650-1671.
Linear and logistic regression was used for birth weight and
infant mortality, respectively. Three models were used for the
analysis: 1) an “empty” random effects model was used to
assess for variation due to clustering within counties, 2) a
fixed effects model assessing for the crude effect of the
county poverty rate on infant mortality and birth weight, and
3) a fixed effects model adjusting for maternal characteristics.
5. Reidpath, D., & Allotey, P. (2003). Infant mortality rate as an indicator of population health.
Journal of Epidemiology and Community Health, 344-346.
2. Lynch, J., & Kaplan, G. (2000). Socioeconomic Position. In L. Berkman, & I. Kawachi,
Social Epidemiology (pp. 13-35). New York, NY: Oxford.
3. Peoples-Sheps, M. D., Guild, P. A., Farel, A. M., Cassady, C. E., Kennelly, J.,
Potrzebowski, P. W., et al. (1998). Model Indicators for Maternal and Child Health: An
Overview of Process, Product, and Applications. Maternal and Child Health Journal, 241-256.
4. Rabe-Hesketh, S., & Skrondal, A. (2012). Multilevel and Longitudinal Modeling using
Stata. College Station, TX: Stata Press.
6. U.S. Census Bureau. (2011). 2006-2010 American Community Survey Kansas.
Acknowledgements
Kansas Department of Health and Environment – Bureau of Epidemiology and Public Health Informatics
Kansas “Region 15” Public Health Preparedness