Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Association of neighborhood characteristics with bystander CPR and out-of-hospital cardiac arrests in the city of Los Angeles
(USC Thesis Other)
Association of neighborhood characteristics with bystander CPR and out-of-hospital cardiac arrests in the city of Los Angeles
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
1
Association of Neighborhood Characteristics with Bystander CPR
and Out-of-Hospital Cardiac Arrests in the City of Los Angeles
by
Wenhao Dong
A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(APPLIED BIOSTATISTICS AND EPIDEMIOLOGY)
May 2018
2
Acknowledgement
I would like to express my deepest gratitude to my master thesis committee chair Dr. Christinne
Lane and all the committee members, Dr. Stephan Sanko and Dr. Meredith Franklin. Without their help
and guidance, I would not finish my master thesis.
I would very much like to express my appreciation to Dr. Lane, for guiding me throughout the
entire PM 594 course, giving me a lot of suggestions on analyzing the data, answering my questions on
thesis with kindness and patience. Additionally, I would not finish this work without Dr. Sanko’s 911
data. He also gave me a lot of help and great suggestions. I would also like to express my appreciation to
Dr. Franklin, who help me a lot with the ArcGIS software and help me revise my essay.
Moreover, I would also like to thank all the faculties of Department of Preventive Medicine at
USC, for giving me this opportunity of this research, as well as IT support and the good academic
environment.
Finally, I want to thank my family, for the financial support and encouragement. Without your
help, I would never have the chance for finishing this great program in USC.
3
Contents
Acknowledgement ........................................................................................................................................ 2
List of Tables ................................................................................................................................................. 4
List of Figures ................................................................................................................................................ 5
Abstract ......................................................................................................................................................... 6
Introduction .................................................................................................................................................. 7
Methods ...................................................................................................................................................... 10
Results ......................................................................................................................................................... 13
Discussion.................................................................................................................................................... 32
Reference .................................................................................................................................................... 35
4
List of Tables
Table 1. Demographic Characteristics of Census Tracts in LAC (Ordered by by-stander CPR rates) 14
Table 2. Demographic Characteristics of Census Tracts in LAC (Ordered by event numbers) 20
Table 3. Frequency table of Bystander CPR and LA-TDS performance 26
Table 4. Pairwise Pearson correlation coefficients 27
Table 5: Census tracts with high OHCA event number and low bystander CPR rate 29
5
List of Figures
Figure 1. The constitute of LA Fire Department attended OHCA events. 9
Figure 2. Age distribution in census tracts of 20% lowest Bystander CPR rate and the other 80% 15
Figure 3. Race distribution in census tracts of 20% lowest Bystander CPR rate and the other 80% 16
Figure 4. Language distribution in census tracts of 20% lowest Bystander CPR rate and the other 80% 16
Figure 5. Education distribution in census tracts of 20% lowest Bystander CPR rate and the other 80% 17
Figure 6. Mean education distribution in census tracts of 20% lowest Bystander CPR rate and the other 80% 18
Figure 7. Income distribution in census tracts of 20% lowest Bystander CPR rate and the other 80% 19
Figure 8. Median income distribution in census tracts of 20% lowest Bystander CPR rate and the other 80% 19
Figure 9. Age distribution in census tracts with the 20% highest OHCA event number and the other 80% census
tracts 21
Figure 10. Race distribution in census tracts with the 20% highest OHCA event number and the other 80% census
tracts 22
Figure 11. Household distribution in census tracts with the 20% highest OHCA event number and the other 80%
census tracts 23
Figure 12. Income distribution in census tracts with the 20% highest OHCA event number and the other 80%
census tracts 24
Figure 13. Median income distribution in census tracts with the 20% highest OHCA event number and the other
80% census tracts
Figure 14. The scatterplot of Bystander CPR rate and event number 25
Figure 15. Census tracts with high OHCA event number and low bystander CPR rate in the City of Los Angeles,
2012-2016 29
Figure 16. Census tracts with different bystander CPR rate in the City of Los Angeles, 2012-2016 30
Figure 17. Census tracts with different OHCA events in the City of Los Angeles, 2012-2016 31
6
Abstract
Early un-interrupted bystander CPR is strongly associated with improved survival from out-of-
hospital cardiac arrest (OHCA); however, different communities within Los Angeles appear to have
different rates of bystander CPR. The Los Angeles Fire Department is interested in measuring and
improving rates of bystander CPR in communities where it is low. Therefore, it is very necessary to
identify target communities within the City of Los Angeles where a public health campaign to promote
bystander CPR would have the greatest impact. This paper describes the communities in the City of Los
Angeles during 2012 to 2016 where the rate of Bystander CPR was the lowest, yet there were a critical
mass of OHCA events. All the census tracts in the City of Los Angeles which had Out-of-hospital cardiac
arrest event during 2012-2016 were classified into the 20% lowest bystander CPR rate group and the
other 80% census tracts group. Characteristics such as gender, age, household, income, education, race
and language were compared between two groups. Using chi-square test, we found that census tracts with
higher proportion of Hispanic people and lower proportion of White people had received less education
and had lower income, and also had lower bystander CPR rates. Similarly, the 1005 census tracts were
classified into the 20% highest bystander CPR rate group and the other 80% census tracts group. The
communities with high OHCA event number had more total population, a slightly larger proportion of
older inhabitants (aged 60 years and older), higher proportion of Black people and lower proportion of
Hispanic people, larger percentage of one person households and lower median income. Besides, after
Los Angeles Tiered-Dispatch System was used, the bystander CPR performance was statistically
significantly improved. The communities with low bystander CPR and high OHCA events were listed in
a table. Some examples that supported our results were provided, and some exclusions were also provided
and analyzed.
Key words: Out-of-hospital cardiac arrest, Bystander CPR, socio-economic characteristics
7
Introduction
Out-of-hospital cardiac arrest (OHCA) affects 360,000 people each year in the United States,
including over 4400 annually in the City of Los Angeles. There are about 295,000 deaths per year due to
cardiac arrest in the US
1
. According to the study of Nichol et.al, the survival rates across the US varied
from city to city, ranging from 7.7% to 39.9%
2
. OHCA survival is associated with numerous factors,
such as patient-, bystander-, event-, community- and EMS system-level factors
3
. But among them, only
few of these are modifiable. Modifiable factors promoting OHCA survival include first call to 911, early
un-interrupted bystander cardiopulmonary resuscitation (CPR), and early defibrillation. According to
previous studies, early un-interrupted bystander CPR is strongly associated with improved survival from
out-of-hospital cardiac arrest (OHCA)
4
.
The neighborhood where a person has had cardiac arrest may also greatly affect his or her
probability of receiving CPR and as a result surviving an OHCA
5
. For example, previous studies
indicated that people who live in neighborhoods that are primarily black, Latino, or poor are more likely
to have an OHCA, less likely to receive CPR, and are less likely to survive
6
. Sasson et.al performed a
large cohort study from 2005 to 2009, and found that patients who had an out-of-hospital cardiac arrest in
low-income black neighborhoods were less likely to receive bystander-initiated CPR than those in high-
income white neighborhoods
7
. Various approaches have been used by other authors to identify
neighborhoods within large cities where bystander CPR rates are low: Besides analyzing the demographic
characteristics of different communities
8, 9
, some studies also used spatial analysis approach to find
significant geographic clusters of low Bystander CPR
10, 11
. Root et.al found two statistically significant
clusters of bystander CPR in Houston and one borderline statistically significant cluster in Austin
12
.
Semple et.al used Local Moran’s I to identify clusters of high OHCA high rates and clusters of low
bystander participation rates in Columbus, Ohio
13
.
In the City of Los Angeles, among all the OHCA events attended by Los Angeles Fire
Department (LAFD), some received resuscitation, while others did not. For those who received
resuscitation, it was either performed by the trained callers, or by someone without any training. Among
these no-training resuscitation, some were witnessed and others were not. What we are interested in this
study are the OHCA cases where resuscitation was attempted by LAFD, no-training, witnessed,
Bystander and No-Bystander CPR (Figure 1).
8
Although the benefit of bystander CPR is well established, previous studies have shown that few
patients in Los Angeles with OHCA receive bystander CPR, including only 13% of cases involving
African American or Latino patients. Moreover, different communities within Los Angeles appear to have
different rates of bystander CPR. Therefore, the overarching goal of this project is to identify target
communities within the City of Los Angeles where a public health campaign to promote bystander CPR
would have the greatest impact. A ‘community’ can be a geographical area, such as areas defined by
different census tracts, or, it can be a sub-population with similar characteristics (like age, race, income or
education). In this study, the communities in the City of Los Angeles were described where the rate of
Bystander CPR was the lowest, and there was a critical mass of OHCA events.
What’s more, in 2014, LAFD developed a new series of scripted questions that drastically
decreases the number of questions needed to identify potential victims of cardiac arrest and lowers the
threshold to provide DA-CPR. This new system, called the Los Angeles Tiered-Dispatch System (LA-
TDS), went into use on December 1, 2014. Since this study was from 2012 to 2016, the LA-TDS
performance is also an important factor that we take into consideration.
9
Figure 1. The constitute of LA Fire Department attended OHCA events. (Resus: Resuscitation. EMS:
Emergency Medical Services. CPR+: CPR performed. CPR-: CPR not performed.)
10
Methods
We analyzed data from the Los Angeles Fire Department (LAFD) Cardiac Arrest Registry from
January 1
st
, 2012 to December 31
st
, 2016. The cases of LAFD-attended OHCA where resuscitation was
attempted were classified into bystander CPR and no bystander CPR.
There were 9090 cases (4558 bystander CPR and 4532 No bystander CPR) records reported
during 2012 to 2016 among 1012 6-digits census tracts in LA County. Five cases were excluded due to
missing census tracts. Three census tracts were excluded, since education, language and income data
were missing (9800.21,9800.22 and 9800.33). Some census tracts were also excluded due to infrequent
number of events (eg. Census tract 9800.09 where is Griffith Park was excluded with event number 11
and total population 14. Census tract 9800.28 where is LAX international airport was excluded with
event number 88 and total population 4. Census tract 3200 where is Universal Studio was excluded with
event number 1 and total population 0. In the end, 1005 census tracts and 8980 events were included in
the final analysis.
To describe the communities in the City of Los Angeles where the rate of Bystander CPR is the
lowest, and there is a critical mass of OHCA events, we firstly calculate the bystander CPR rate in each
census tract (bystander CPR number/event number). Then we ordered the data by ascending bystander
CPR rate, comparing the lowest 20 percent (201) census tracts characteristics with the other 80 percent
census tracts (804). The data was also ordered by descending event number, comparing the highest 20
percent (201) census tracts characteristics with the other 80 percent census tracts (804).
Characteristics that were used for further analysis included gender, age, race, household status,
education, language and income. Age had been categorized into five groups (≤9, 10-19, 20-34, 35-59,
≥60). Race was categorized into four groups (Hispanic, Black, White and other). Household status was
categorized into three groups (one person household, family which has two or more people, non-family
which has two or more people). Education was categorized into three groups (primary school or less,
middle to high school, higher education). Education was also treated as a continuous variable, mean
education year, which was calculated using total education years divided by total population. Language
was categorized into three groups (English speaking only, Spanish or Spanish Creole speaking and other
language speaking). Yearly income level was categorized into three groups (<35k, 35-73k, >75k). It was
also treated as a continuous variable using mean and median income. The proportions of each gender, age
11
group, race, household status, education level, language group and income level were calculated within
each census tract (number on each category/total population*100). In 2014, LA Fire Department
developed a new card system, called the Los Angeles Tiered-Dispatch System (LA-TDS), and it went into
use on December 1, 2014. For the LAFD Tiered-Dispatch System (LA-TDS) effect, we set indicator for
before and within/after Dec of 2014.
To describe the community characteristics in the City of Los Angeles where the rate of bystander
CPR rate is the lowest, we first calculated the bystander CPR performance rate in each census tract (the
number of bystander CPR/ total event number), then ordered all of the census tracts by ascending
bystander CPR rate. The census tracts with the lowest 20% (201) bystander CPR rate were classified as
low bystander CPR rate group, and the other 80% (804) were classified as the other group. The averages
of the variables needed (the rate of each gender, age group, race, household status, education level, mean
education year, language, income level, mean and median income) were calculated separately in each
group. For categorical variables, a frequency table is created with N (%). And continuous characteristics
are described by mean and standard deviation. Histograms and boxplots were also used to describe the
demographic characters, histogram for category variables and boxplots for continuous variables.
To compare the characteristics of low bystander CPR rate group with the other group, we used
chi-square test to explore the difference of category variables (eg. gender, age, race, household status,
language, education levels and income levels), and Student’s t-test to explore the difference of continuous
variables (eg. mean education year, median income and mean income). To explore the effect of LATDS
system on bystander CPR performance, 2*2 table with frequency was created and chi-square test was
used to test the significance.
Similarly, to describe the community characteristics in City of Los Angeles where the OHCA
events is the highest, we firstly counted the event number in each census tract, then order all the census
tracts by descending OHCA event number. The census tracts with the highest 20% (201) event number
were classified as high event number group, and the other 80% (804) were classified as the other group.
All the characteristics were analyzed the same way with above.
Then we tested the association between each two variables using Pearson’s correlation
coefficient. To find the communities in the City of Los Angeles where the rate of Bystander CPR is the
lowest, and there is a critical mass of OHCA events, we also used Pearson’s correlation coefficient and
draw the scatterplot of bystander CPR rate and OHCA event number.
To more clearly see the communities in the City of Los Angeles where the rate of Bystander CPR
was the lowest, and there was a critical mass of OHCA events more straightforward, ArcMap 10.5.1
12
software was used to draw maps of bystander CPR rate and OHCA event numbers in the City of Los
Angeles. The ‘Geography’ variable was changed to standard census tract format. Then we classified the
bystander CPR rate into 5 groups: 0-0.263, 0.263-0.4, 0.4-0.5,0.5-0.667,0.667-1; and OHCA event
number into 5 groups: 1-3, 3-5,5-8,8-12,12-84; all the cut-points are quintiles. Graduated colors were
used to represent the quantities.
All statistical analyses were conducted with the use of SAS software, version 9.2. P values are
based on a two-sided significance level of 0.05.
13
Results
A total of 8980 cases within 1005 census tracts met the criteria for inclusion in our analyses. The
communities having the 20% lowest bystander CPR rate ranged from 0 to 0.263, while the other 80%
ranged from 0.267 to 1. A description of the community characteristics for the two bystander CPR rate
groups can be found in Table 1.
14
Examining the two bystander CPR groups by sex, we found that there was no statistically
significant difference between them (p=0.944).
Similarly, bystander CPR rates for different age groups were almost the same (p=0.991). When
we did the multiple comparison, we found that in some age categories, there were statistically significant
differences between the two groups: p=0.004 in less than 9 years old category, p=0.045 in 10-19 years
15
old category, p<0.001 in over 60 years old category. However, the absolute values did not have great
differences (Figure 2).
Figure 2. Age distribution in census tracts of 20% lowest Bystander CPR rate and the other 80%
When examined by race, we found there were no statistically significant differences of the overall
race distribution between two groups (p=0.539). However, there was a greater percentage of Hispanic
people in the communities of low bystander CPR rate: 54.73% in low bystander CPR rate group and
46.01% in the other group (p<0.001). Also, the proportion of white people was lower in the low bystander
CPR rate group: 22.8% versus 31.22% (p<0.001) (Figure 3).
16
Figure 3. Race distribution in census tracts of 20% lowest Bystander CPR rate and the other 80%
There were no statistically significant differences of the overall language distribution between
two groups (p=0.539). However, there were less percentage of people who speak English only in the
communities of low bystander CPR rate: 35.11% in low bystander CPR rate group and 41.48% in the
other group (p<0.001). Also, there were larger proportion of people who speak Spanish or Spanish Creole
in the communities of low bystander CPR rate: 49.5% in low bystander CPR rate group and 41.17% in
the other group (p<0.001) (Figure 4).
Figure 4. Language distribution in census tracts of 20% lowest Bystander CPR rate and the other 80%
17
There was no statistically significant difference of the household distribution between low
bystander CPR rate group and the other one (p=0.997).
When the education was treated as a categorical variable and classified into three groups (primary
school or less, middle to high school and higher education), there was no statistically significant
difference of education distribution between low bystander CPR rate group and the other (p=0.726). But
when comparing the percentage in each category, we found that there were less percentage of people who
had received higher education in communities with low bystander CPR rate: 66.35% in low bystander
CPR rate group and 73.13% in the other (p<0.001) (Figure 5).
Figure 5. Education distribution in census tracts of 20% lowest Bystander CPR rate and the other 80%
When we treated education as a continuous variable and calculated the mean education year by
groups, we found that the average mean education year in low bystander CPR group was statistically
significant lower than that of the other group (p<0.001). People in communities with low bystander CPR
rate had a mean education of 11.67 years, while 12.43 years in the other communities (Figure 6).
18
Figure 6. Mean education distribution in census tracts of 20% lowest Bystander CPR rate and the other 80%.
Group 1 represents the census tracts with 20% lowest CPR rate, and Group 2 represents the other 80% census
tracts
According to the estimate released by the U.S. Department of Housing and Urban Development
(HUD), the household income below $31,550 would be considered to a very low income. Also, the
median family income for Los Angeles County is determined by HUD to be $64,300. Therefore, we
categorized the household income into three groups: below 35k per year, 35k to 75k per year and above
75k per year. There were more percentage of households which had annual income below 35k in low
bystander CPR group (43.7%) than in the other group (37.31%) (p<0.001). And there were less
percentage of households which had annual income above 75k in low bystander CPR group (24.93%)
than in the other group (32.52%) (p<0.001) (Figure 7).
19
Figure 7. Income distribution in census tracts of 20% lowest Bystander CPR rate and the other 80%
Since the family income is not normally distributed in the City of LA (there are many people who
have extremely high income and people who have very low income), we used median income instead of
mean income in the analysis. In the boxplot, group 1 represents the low bystander CPR rate group and
group 2 represents the other one. We can find that the average median income in group 1 was statistically
significant lower than that of group 2. People in communities with low bystander CPR rate had an
average annual median income of 45.37k, while 57.39k in the other communities (p<0.001) (Figure 8).
Figure 8. Median income distribution in census tracts of 20% lowest Bystander CPR rate and the other 80%. Group
1 represents the census tracts with 20% lowest CPR rate, and Group 2 represents the other 80% census tracts.
20
Similarly, to describe the communities in the City of Los Angeles where there is a critical mass of
OHCA events, all the demographic characteristics of high bystander CPR event number group and the
other group were described and compared in table 2. The highest 20% event number ranged from 84 to
12, and the other 80% ranged from 1 to 12.
21
Examining the two bystander CPR groups by sex, we found that there was no statistically
significant difference between them (p=0.991).
The total population was statistically significant higher in the communities with high OHCA
event number than in the other group (p<0.001).
The age distributions were almost the same (p=0.991) between the two groups (Figure 9). The
communities of high OHCA event number had a slightly larger proportion of older inhabitants: there were
16.91% of people aged 60 years and older, and 14.88% in the other areas (p<0.001).
Figure 9. Age distribution in census tracts with the 20% highest OHCA event number and the other 80% census
tracts.
There were no statistically significant differences of the overall distribution between two groups
(p=0.239). However, there were less percentage of Hispanic people in the communities of high OHCA
event number: 42.58% in high OHCA event number group and 49.07% in the other group (p<0.001).
Also, the proportion of black people was higher in the high OHCA event number group, which was
16.42%, and only 7.16% in the other group (p<0.001). Chi-square test was used to see if the proportion of
Hispanic and Black people were different between groups, the result showed that there was statistically
significant difference of Hispanic and Black distribution between the two groups (p=0.045) (Figure 10).
22
Figure 10. Race distribution in census tracts with the 20% highest OHCA event number and the other 80% census
tracts.
There was no statistically significant difference of the household distribution between high
OHCA event number group and the other one (p=0.821). But when comparing the percentage in each
category, we found that there were more percentage of people who lived by themselves in communities
with high OHCA event number: 28.47% in high OHCA event number group and 24.96% in the
other(p<0.001). And there were less percentage of people who live with 2 or more people (71.53%) in
high OHCA event number group than that of the other (75.04%) (p<0.001) (Figure 11).
23
Figure 11. Household distribution in census tracts with the 20% highest OHCA event number and the other 80%
census tracts.
When the education was treated as a category variable and classified into three groups (primary
school or less, middle to high school and higher education), there was no statistically significant
difference of education distribution between high OHCA event number group and the other (p=0.949).
When we treated education as a continuous variable and calculated the mean education year by
groups, we found that there was no significant difference of the average mean education year between
two groups (p=0.864). People in communities with high OHCA event number had a mean education of
12.30 years, while 12.27 years in the other communities.
For the same reason we discussed above, the household income was categorized into three
groups: below 35k per year, 35k to 75k per year and above 75k per year. There was no overall significant
difference between the two groups (p=0.892). But when comparing the percentage in each category, there
were greater percentage of households which had annual income below 35k in high OHCA event number
group (41.08%) than in the other group (37.99%) (p<0.001). And there were slightly less percentage of
households which had annual income above 75k in high OHCA event number group (28.99%) than in the
other group (31.48%) (p<0.001) (Figure 12).
24
Figure 12. Income distribution in census tracts with the 20% highest OHCA event number and the other 80%
census tracts
Because of the same reason discussed above, we used median income instead of mean income in
the analysis. In the boxplot, group 1 represents the high OHCA event number group and group 2
represents the other one. We can find that the average median income in group 1 was marginally
significant lower than that of group 2 (p=0.056). People in communities with high OHCA event number
had an average annual median income of 51.19k, while 57.39k in the other communities (Figure 13).
25
Figure 13. Median income distribution in census tracts with the 20% highest OHCA event number and the other
80% census tracts. Group 1 represents the census tracts with 20% highest OHCA events, and Group 2 represents
the other 80% census tracts.
To explore the effect of LATDS system on bystander CPR performance, all the OHCA events
from 2012 to 2016 were classified into four groups: with or without LATDS, and performed or not
performed bystander CPR. Two by two table with frequency of each group was created and chi-square
test was used to test the significance. Boxplot was also drawn to compare the groups with or without
LATDS. From Table 3, we could see that before the LATDS system was implemented, the percentage of
bystander CPR performance was only about 48%. While after LATDS was used, this percentage
increased to 52.5%. The Chi-square 1-sided p-value was less than 0.0001, which meant after LATDS was
used, the bystander CPR performance was statistically significantly improved.
26
Table 3. Frequency table of Bystander CPR and LA-TDS performance.
To identifying the characteristics of areas with low bystander CPR rates more deeply, we also
tested the collinearity between the characteristics. After excluded missing values, there were 1001 census
tracts included in the final analysis. The result showed that there was a statistically significant positive
linear correlation between the proportion of Hispanic people and people who speak Spanish (p<0.001).
The correlation coefficient (R) is 0.978, which means a very strong correlation. There was a statistically
significant negative linear correlation between the proportion of Hispanic people and mean education year
(p<0.001), and they were strongly negatively correlated (R=-0.917). The proportion of White people was
statistically significantly positively correlated with mean education year (p<0.001, R=0.857). And it was
also statistically significantly positively correlated with median income (p<0.001, R=0.734) and people
who speak English only (p<0.001, R=0.769). What’s more, there was statistically significant association
between mean education year and median income (p<0.001), and they were strongly positively correlated
(R=0.712).
Bystander CPR
System Not Performed(%) Performed(%) Total
No LA-TDS 2533 (52.10%) 2329 (47.90%) 4862
LA-TDS 1957 (47.52%) 2161 (52.48%) 4118
Total 4490 4490 8980
27
Table 4. Pairwise Pearson correlation coefficients
Pairwise Pearson correlation coefficients (N=1001)
Hispanic Hispanic
1.00000
Spanish
speaking
0.97816
<0.0001
Mean
education year
-0.91685
<0.0001
English
speaking
-0.85861
<0.0001
White
-0.84820
<0.0001
White White
1.0000
Mean
education year
0.85679
<0.0001
Hispanic
-0.84820
<0.0001
Spanish
speaking
-0.84069
<0.0001
English
speaking
0.76866
<0.0001
English English speaking
1.0000
Spanish
speaking
-0.88505
<0.0001
Hispanic
-0.85861
<0.0001
Mean
education year
0.83298
<0.0001
White
0.76866
<0.0001
Spanish Spanish speaking
1.0000
Hispanic
0.97816
<0.0001
Mean
education year
-0.91810
<0.0001
English
speaking
-0.88505
<0.0001
White
-0.84069
<0.0001
Mean
education
year
Mean education
year
1.0000
Spanish
speaking
-0.91810
<0.0001
Hispanic
-0.91685
<0.0001
White
0.85679
<0.0001
English
speaking
0.83298
<0.0001
Median
income
Median income
1.0000
White
0.73499
<0.0001
Mean
education year
0.71159
<0.0001
Spanish
speaking
-0.63525
<0.0001
English
speaking
0.63425
<0.0001
28
Figure 14. The scatterplot of Bystander CPR rate and event number. Each circle represents a census tract. Two
reference lines are added (x=12, y=0.263) and divide all the circles into four parts. The circles distributed on the
right below zone (which is marked by red line) are what we interested in this study.
OHCA incidence and CPR prevalence had a modest correlation with a correlation coefficient of
0.143. But this correlation was statistically significant (P<0.001) by Pearson Correlation Coefficient Test.
From the scatterplot of bystander CPR rate and OHCA event number, it is more straight forward to find
the communities in the City of Los Angeles where the rate of Bystander CPR is the lowest, and there is a
critical mass of OHCA events. Two reference lines were drawn on the plot: bystander CPR rate equaled
to 0.263 (the lowest 20%) and event number equaled to 12 (the highest 20%). Each circle represents a
census tract in City of Los Angeles, and those drop into the zone with red line are what we are interested
(bystander CPR rate less than 0.263 and event number more than 12). All the areas were listed in the table
5. There were total 26 census tracts which we are interested to promote bystander CPR there. To see these
areas more straightly and clearly, maps were drawn using ArcGIS software, the census tracts with high
OHCA event number and low bystander CPR rate were marked by red (Figure 14).
29
Table 5: Census tracts with high OHCA event number and low bystander CPR rate
Figure 15. Census tracts with high OHCA event number and low bystander CPR rate
in the City of Los Angeles, 2012-2016
30
We also drew the maps of census tracts with different bystander CPR rate and OHCA event
number in the City of Los Angeles, 2012-2016. Graduated colors were used to represent the quantities.
We could see that in some census tracts of Santa Monica, Venice, Burbank and Pasadena cities, the
bystander CPR rates were high and the OHCA events were low; while in some census tracts of downtown
Los Angeles and Boyle Heights, the bystander CPR rates were low and the OHCA events were high.
Figure 16. Census tracts with different bystander CPR rate in the City of Los Angeles, 2012-2016
31
Figure 17. Census tracts with different OHCA events in the City of Los Angeles, 2012-2016
32
Discussion
In this study, in cases of 911-attended OHCA where resuscitation was attempted during 2012-
2016 in the City of Los Angeles, some community characteristics that were mostly associated with low
Bystander CPR rate and high OHCA events have been identified. The communities with low Bystander
CPR rate had higher proportion of Hispanic and lower proportion of White; larger percentage of people
who speak Spanish or Spanish Creole, and less percentage of people who speak English only (this is
consistent with the race characteristic, since Hispanic people mostly speak Spanish and White people
mostly speak English only). There were fewer proportion of people who had received higher education in
the neighborhood of low Bystander CPR rate, and the mean education year was also lower there, which
means people receive less education may less likely to perform bystander CPR. Moreover, in the areas
with low Bystander CPR rate, larger percentage of people had annual family income below the poverty
line of Los Angeles, and the family annual median income was lower. In conclusion, the racial
composition and socioeconomic factors (education and income) of neighborhood are most associated with
the likelihood of receiving bystander CPR in the City of Los Angeles. We can also learn from the
correlation test that the mean education year and median income had statistically significant negative
association with the proportion of Hispanic people, and had significant negative association with the
proportion of White people. All these results indicated that communities with higher proportion of
Hispanic people and lower proportion of White people had receive less education and had lower income,
therefore had lower bystander CPR rate, and vise versa. Besides, after LATDS was used, the bystander
CPR performance was statistically significantly improved. The communities with high OHCA event
number had more total population, a slightly larger proportion of older inhabitants (aged 60 years and
older), higher proportion of Black people and lower proportion of Hispanic people. What’s more, there
was larger percentage of one person households and lower median income in the areas of high OHCA
events.
These results above are consistent with some previous studies. Sasson et. al found a direct
relationship between the median income and racial composition of a neighborhood and the probability
that a person with out-of-hospital cardiac arrest received bystander-initiated CPR
7
. In this study, persons
with cardiac arrest in a high-income white neighborhood were the most likely to receive bystander-
initiated CPR in the 29 sites of CARE. The researches from Canada
14
and Seattle
14
indicated that the
OHCA events which happened in an area of higher socioeconomic status, which was associated with
higher education, were more likely to receive CPR. Besides, Sasson et.al pointed out that Black and
Latino adults were more likely than white adults to have an out-of-hospital cardiac arrest (OHCA) in
33
Columbus, Ohio
15
. The study of Franklin County, Ohio demonstrated that compared to the entire county,
the high-risk communities of unattended OHCA had a slightly larger proportion of older inhabitants,
lower median household income and a larger percentage of inhabitants living below the poverty level
13
.
However, there are also some researches that are not consistent with our results. For instance, data from
Chicago in the 1980s suggested that the racial composition of a neighborhood was the only important
predictor of the likelihood of receiving bystander CPR, instead of median income and education
attainment
5
.
A number of census tracts have also been identified where should be targeting with a public
campaign to promote bystander CPR and improve OHCA outcomes in the future. These census tracts all
had low bystander CPR rate as well as high OHCA events (Table 5). After exploring the neighborhood
characteristics of these areas, we found that most of them were consistent with our findings. For example,
Census Tract 2260.02 in Los Angeles County, California (Boyle Heights), with an event number of 20
and Bystander CPR rate 0.25 during 2012-2016, has a high proportion of Hispanic people (49.81%), a
larger percentage of inhabitants living below the poverty level (67%) and a low median income (23.6k).
Another example is that Census Tract 2063, Los Angeles County, California (Boyles Heights). It had an
event number of 49 and bystander CPR rate of 0.2. The racial composition in this community is mostly
Black (52.82%) and the mean education year is low (11.38). Most of inhabitants have family income
below the poverty line (96.7%) and live by themselves (94.74% one person household). One more unique
characteristics of this communities which we have not found above is that there are much more male than
female (75.5% vs 24.5%). Since the sex of almost all the communities in LA are distributed evenly, we
did not find any association between the sex composition and bystander CPR performance, similar to
other studies
7, 8, 12
. Census Tract 2063’s example indicated that the association may actually exist and need
further study in the future. Another interesting example is Census Tract 2735.02, Los Angeles County,
California (Venice), which had an event number of 16 and a bystander CPR rate 0.25. Most inhabitants in
this community are White (77.98%). The mean education year is as much as 15.62 and a large proportion
of people have received higher education (84.99%). According to previous results, communities like this
should have low bystander CPR rate and low risk of unattended OHCA events, but Census Tract 2735.02
became an exclusion. I think the reason maybe the composition of households, the percentage of 1 person
households in this community is 57.79%, which is much higher than the average level. So besides the
racial composition and socioeconomic characteristics, the proportion of people living alone is also an
important factor that should be paid attention.
In conclusion, we found that the racial and socio-economic composition of neighborhoods has
important effects on the likelihood of bystander CPR for a person with an OHCA; racial, household and
34
income composition are mostly associated with OHCA events. These results are similar to many previous
researches in other areas, indicating the reliability of our study. This finding suggests that CPR training
targeted to neighborhoods with racial and socioeconomic characteristics associated with a low probability
of bystander CPR may constitute an evidence-based approach to public health planning.
35
Reference
[1] Roger VL, Go AS, Lloyd-Jones DM, Benjamin EJ, Berry JD, Borden WB, Bravata DM, Dai S, Ford ES, Fox
CS, Fullerton HJ, Gillespie C, Hailpern SM, Heit JA, Howard VJ, Kissela BM, Kittner SJ, Lackland DT,
Lichtman JH, Lisabeth LD, Makuc DM, Marcus GM, Marelli A, Matchar DB, Moy CS, Mozaffarian D,
Mussolino ME, Nichol G, Paynter NP, Soliman EZ, Sorlie PD, Sotoodehnia N, Turan TN, Virani SS, Wong
ND, Woo D, Turner MB, American Heart Association Statistics C, Stroke Statistics S: Heart disease and
stroke statistics--2012 update: a report from the American Heart Association. Circulation 2012, 125:e2-
e220.
[2] Nichol G, Thomas E, Callaway CW, Hedges J, Powell JL, Aufderheide TP, Rea T, Lowe R, Brown T,
Dreyer J, Davis D, Idris A, Stiell I, Resuscitation Outcomes Consortium I: Regional variation in out-of-
hospital cardiac arrest incidence and outcome. JAMA 2008, 300:1423-31.
[3] Strategies to Improve Cardiac Arrest Survival: A Time to Act. Edited by Graham R, McCoy MA, Schultz
AM. Washington (DC), 2015.
[4] Fordyce CB, Hansen CM, Kragholm K, Dupre ME, Jollis JG, Roettig ML, Becker LB, Hansen SM,
Hinohara TT, Corbett CC, Monk L, Nelson RD, Pearson DA, Tyson C, van Diepen S, Anderson ML, McNally
B, Granger CB: Association of Public Health Initiatives With Outcomes for Out-of-Hospital Cardiac Arrest
at Home and in Public Locations. JAMA Cardiol 2017.
[5] Iwashyna TJ, Christakis NA, Becker LB: Neighborhoods matter: a population-based study of provision
of cardiopulmonary resuscitation. Ann Emerg Med 1999, 34:459-68.
[6] Galea S, Blaney S, Nandi A, Silverman R, Vlahov D, Foltin G, Kusick M, Tunik M, Richmond N:
Explaining racial disparities in incidence of and survival from out-of-hospital cardiac arrest. Am J
Epidemiol 2007, 166:534-43.
[7] Sasson C, Magid DJ, Chan P, Root ED, McNally BF, Kellermann AL, Haukoos JS, Group CS: Association
of neighborhood characteristics with bystander-initiated CPR. N Engl J Med 2012, 367:1607-15.
[8] Johnson MA, Grahan BJH, Haukoos JS, McNally B, Campbell R, Sasson C, Slattery DE: Demographics,
bystander CPR, and AED use in out-of-hospital pediatric arrests. Resuscitation 2014, 85:920-6.
[9] Becker LB, Han BH, Meyer PM, Wright FA, Rhodes KV, Smith DW, Barrett J: Racial differences in the
incidence of cardiac arrest and subsequent survival. The CPR Chicago Project. N Engl J Med 1993,
329:600-6.
[10] Sasson C, Cudnik MT, Nassel A, Semple H, Magid DJ, Sayre M, Keseg D, Haukoos JS, Warden CR,
Columbus Study G: Identifying high-risk geographic areas for cardiac arrest using three methods for
cluster analysis. Acad Emerg Med 2012, 19:139-46.
[11] Warden C, Cudnik MT, Sasson C, Schwartz G, Semple H: Poisson cluster analysis of cardiac arrest
incidence in Columbus, Ohio. Prehosp Emerg Care 2012, 16:338-46.
[12] Root ED, Gonzales L, Persse DE, Hinchey PR, McNally B, Sasson C: A tale of two cities: the role of
neighborhood socioeconomic status in spatial clustering of bystander CPR in Austin and Houston.
Resuscitation 2013, 84:752-9.
[13] Semple HM, Cudnik MT, Sayre M, Keseg D, Warden CR, Sasson C, Columbus Study G: Identification
of high-risk communities for unattended out-of-hospital cardiac arrests using GIS. J Community Health
2013, 38:277-84.
[14] Vaillancourt C, Lui A, De Maio VJ, Wells GA, Stiell IG: Socioeconomic status influences bystander CPR
and survival rates for out-of-hospital cardiac arrest victims. Resuscitation 2008, 79:417-23.
36
[15] Sasson C, Haukoos JS, Bond C, Rabe M, Colbert SH, King R, Sayre M, Heisler M: Barriers and
facilitators to learning and performing cardiopulmonary resuscitation in neighborhoods with low
bystander cardiopulmonary resuscitation prevalence and high rates of cardiac arrest in Columbus, OH.
Circ Cardiovasc Qual Outcomes 2013, 6:550-8.
Abstract (if available)
Abstract
Early un-interrupted bystander CPR is strongly associated with improved survival from out-of-hospital cardiac arrest (OHCA)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Evaluation of factors influencing Los Angeles Tiered-Dispatch System’s improvement on bystander CPR rate and inter reliability between electronic patient care report (ePCR) and 911 call review on...
PDF
Spatial analysis of PM₂.₅ air pollution in association with hospital admissions in California
PDF
Characteristics associated with emergency medical services for cardiac arrest pre- and during COVID-19 in Los Angeles, CA
PDF
Contemporary outcomes for adult congenital heart surgery in an adult tertiary care hospital
PDF
Predictive factors of breast cancer survival: a population-based study
PDF
Spatiotemporal studies of out-of-hospital cardiac arrests and bystander cardiopulmonary resuscitation in Los Angeles
PDF
Assessment of the mortality burden associated with ambient air pollution in rural and urban areas of India
PDF
Street connectivity and childhood obesity: a longitudinal, multilevel analysis
PDF
Surgical aortic arch intervention at the time of extended ascending aortic replacement is associated with increased mortality
PDF
A randomized controlled trial of hormone therapy on postmenopausal women’s quality of life
PDF
Evaluating the use of friend or family controls in epidemiologic case-control studies
PDF
Factors associated with agreement between childhood cancer survivors (CCS) and their parents on knowledge and attitudes about health-related quality of life, treatment and follow-up care
PDF
Association of comorbidity with prostate cancer tumor characteristics in African American men
PDF
Associations between isoflavone soy protein (ISP) supplementation and breast cancer in postmenopausal women in the Women’s Isoflavone Soy Health (WISH) clinical trial
PDF
Exploring the association of number of cigarettes smoked and confidence to quit smoking in Korean American emerging adults: a multilevel modeling approach
PDF
Spatial modeling of non-tailpipe emissions and its association with children's lung function
PDF
A cohort study of air-pollution and childhood obesity incidence
PDF
Visual acuity outcomes after cataract extraction in Chinese Americans: the Chinese American Eye Study (CHES)
PDF
A novel risk-based treatment strategy evaluated in pediatric head and neck non-rhabdomyosarcoma soft tissue sarcomas (NRSTS) patients: a survival analysis from the Children's Oncology Group study...
PDF
Carboplatin and vincristine chemotherapy for progressive low grade gliomas in pediatric patients with or without neurofibromatosis type 1 (NF1)
Asset Metadata
Creator
Dong, Wenhao
(author)
Core Title
Association of neighborhood characteristics with bystander CPR and out-of-hospital cardiac arrests in the city of Los Angeles
School
Keck School of Medicine
Degree
Master of Science
Degree Program
Applied Biostatistics and Epidemiology
Publication Date
01/19/2018
Defense Date
01/18/2018
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,out-of-hospital cardiac arrest, bystander CPR, socio-economic characteristics
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Lane, Christinne (
committee chair
), Franklin, Meredith (
committee member
), Sanko, Stephen (
committee member
)
Creator Email
912804971@qq.com,wenhaodo@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c40-464770
Unique identifier
UC11266879
Identifier
etd-DongWenhao-5964.pdf (filename),usctheses-c40-464770 (legacy record id)
Legacy Identifier
etd-DongWenhao-5964.pdf
Dmrecord
464770
Document Type
Thesis
Rights
Dong, Wenhao
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
out-of-hospital cardiac arrest, bystander CPR, socio-economic characteristics