The Association Between Poverty and Sudden Infant Death Syndrome SIDS: United States 1990 and 2000

1
The Association Between Poverty and Sudden Infant
Death Syndrome (SIDS)United States 1990 and 2000
TABLE 2. ODDS RATIOS FOR SIDS AND OTHER CAUSES OF
DEATH BY YEAR AND RACE FOR POVERTY QUARTILES
Michael H. Malloy, Karl Eschbach and Daniel H.
Freeman Departments of Pediatrics, Internal
Medicine/Geriatrics and Office of Biostatistics,
The University of Texas Medical Branch,
Galveston, Texas
Introduction SIDS has been associated with a
number of demographic characteristics that
suggest more economically disadvantaged
populations are at higher risk. Because of the
absence of economic data on U.S. vital statistics
files, no direct measure of association between
economic disadvantage and SIDS has been possible.
TABLE 1 EXAMPLE OF DATA FILE INPUT FOR STATE OF
ALABAMA Obs State Year
Race SIDS Count PVNDX PVQTL
1 Alabama 1990 Hispanic
0 269 0.19368 2
2 Alabama 1990 Hispanic 2
3 0.19368 2 3
Alabama 1990 Black 0 21499
0.37682 0 4 Alabama
1990 Black 1 53
0.37682 0 5 Alabama
1990 Black 2 294
0.37682 0 6 Alabama
1990 White 0 41039
0.11736 0 7 Alabama
1990 White 1 47
0.11736 0 8 Alabama
1990 White 2 283
0.11736 0 9 Alabama
2000 Hispanic 0 1669
0.25047 1 10 Alabama 2000
Hispanic 2 7 0.25047
1 11 Alabama 2000
Black 0 20214 0.31249
0 12 Alabama 2000 Black
1 22 0.31249 0
13 Alabama 2000 Black
2 295 0.31249 0
14 Alabama 2000 White 0
40878 0.10414 1 15
Alabama 2000 White 1 34
0.10414 1 16
Alabama 2000 White 2 244
0.10414 1 SIDS Code0Survivor
1SIDS Victim 2Other Cause of Death
Objective The objective of this ecologic
analysis was to merge Census data with Vital
Statistics data in order to explore associations
between economic disadvantage and the occurrence
of SIDS.
Methods The percent of each states
population below established Federal poverty
guidelines (poverty index-PVNDX) was obtained
from U.S. Census data for 1990 and 2000 by race
(Hispanic-HISP, non-Hispanic white-NHW, and
non-Hispanic black-NHB). These data were
merged by year of birth, state, and race with
U.S. Vital Statistics Linked Birth and Infant
Death Certificate data. Associations between
PVNDX and category of infant outcome (Survivor,
SIDS, and Other Causes of Death). The PVNDX was
analyzed as a nominal variable by dividing it
into quartiles by race. An example of the
appearance of data from one state is provided in
Table 1. Logistic regression modeling was used
to examine the relationship between infant
outcome and poverty.
Results
1. There were 7,803 SIDS and 56,829 other causes
of deaths (OCD) that occurred in the combined
years of 1990 and 2000. 2. There were significant
interactions between year of death and SIDS and
between race and SIDS, thus data are analyzed and
presented by year and race. 3. Figures 1 and 2
demonstrate the distribution of SIDS rates by
state for the years 1990 and 2000. Figures 3
and 4 demonstrate the distribution of poverty
index for each state. Notice that there appears
to be little relationship between level of
poverty index and SIDS rate when data are
presented without consideration of race. 4. Table
2 shows the odds ratio for SIDS and OCD by race
and year. Notice the dose response relationship
between poverty quartile and SIDS for NHW and
NHB, but not for HISP. In addition notice the
much weaker relationship between poverty quartile
and OCD for all races. 5. There was no
significant relationship between the change in
poverty Index and the change in SIDS rate between
1990 and 2000.
Conclusions
1.There does appear to be an association between
poverty and SIDS for NHW and NHB. 2. The lack of
association between poverty and HISP SIDS rates
is consistent with the previously reported
Hispanic paradox. 3. Further analysis of the
impact of poverty at an ecologic level closer to
the level of case ascertainment, i.e. county or
census track level appears warranted.