Abstract
Background
For individuals with an out‐of‐hospital cardiac arrest (OHCA), survival may be influenced by the neighborhood in which the arrest occurs.
Methods and Results
Within the national CARES (Cardiac Arrest Registry to Enhance Survival) registry, we identified 169 502 patients with OHCA from 2013 to 2017. On the basis of census tract data, OHCAs were categorized as occurring in predominantly white (>80% white), majority black (>50% black), or integrated (neither of these 2) neighborhoods and in low‐income (median household <$40 000), middle‐income ($40 000 to $80 000), or high‐income (>$80 000) neighborhoods. With hierarchical logistic regression, the association of neighborhood race and income on overall survival was assessed. Overall, 37.5%, 16.6%, and 45.9% of people had an OHCA in predominantly white, majority black, and integrated neighborhoods, and 30.1%, 53.4%, and 16.5% in low‐, middle‐, and high‐income neighborhoods, respectively. Compared with OHCAs occurring in predominantly white neighborhoods, those in majority black neighborhoods were 12% less likely (6.9% versus 10.6%; adjusted odds ratio 0.88; 95% CI 0.82‐0.95; P<0.001) to survive to discharge, whereas those in integrated neighborhoods had similar survival (10.3% versus 10.6%; adjusted odds ratio 1.00; 95% CI 0.96‐1.04; P=0.93). Compared with high‐income neighborhoods, those in middle‐income neighborhoods were 11% (10.1% versus 11.3%; adjusted odds ratio 0.89; 95% CI 0.8‐0.94; P<0.001) less likely to survive to discharge, whereas those in low‐income neighborhoods were 12% (8.6% versus 11.3%; adjusted odds ratio 95% CI 0.83‐0.94; P<0.001) less likely to survive. Differential rates of bystander cardiopulmonary resuscitation only modestly attenuated neighborhood differences in survival.
Conclusions
OHCAs in majority black and non–high‐income neighborhoods have lower survival rates, and these differences were not explained by differential bystander cardiopulmonary resuscitation rates.
Keywords: cardiac arrest, income, race, survival
Subject Categories: Cardiopulmonary Arrest, Quality and Outcomes
Clinical Perspective
What Is New?
Patients with an out‐of‐hospital cardiac arrest in majority black and non–high‐income neighborhoods have lower overall survival rates, and survival differences are not explained by neighborhood differences in bystander CPR rates.
Although there were differences in overall survival by neighborhood race and income, survival rates to hospital admission were generally similar across neighborhoods.
What Are the Clinical Implications?
Differences in other aspects of emergency medical service response that differ by neighborhood race and income may be important to address in order to reduce out‐of‐hospital cardiac arrest survival disparities.
The lack of neighborhood differences in rates of survival to hospital admission implies that overall survival differences may also be due to differences in postresuscitation care at hospitals.
Our findings suggest the need to improve not only bystander CPR rates but also other aspects of emergency medical service treatment and postresuscitation care in communities that are majority black and non–high income.
Introduction
Out‐of‐hospital cardiac arrest affects ≈350 000 individuals annually in the United States and is a major public health condition.1 Although survival rates are low (<10%),2 there is substantial variation across communities, with more than a 5‐fold difference (3% to 16%) in 1 multicenter registry.3 The influence of neighborhood factors, especially race and income, on one's likelihood of surviving an out‐of‐hospital cardiac arrest, however, is less clear. Further elucidating potential disparities in survival between communities with different racial and income compositions may guide public health strategies to improve bystander education, emergency medical service (EMS) response, and hospital care.
Prior studies of neighborhood effects on out‐of‐hospital cardiac arrest outcomes have been limited but suggest that neighborhood race and socioeconomic status may be important determinants.4, 5, 6, 7, 8, 9 One prior study reported substantially lower rates of bystander cardiopulmonary resuscitation (CPR) in communities that were largely black or low‐income communities, but did not examine neighborhood effects on rates of survival.4 A more recent study from the Resuscitation Outcomes Consortium reported lower rates of bystander CPR, survival to discharge, and other survival outcomes in largely black communities, but neighborhood‐level differences in survival were eliminated after accounting for the fact that most of the black neighborhoods were clustered in 4 of the 10 registry sites.5 Thus, the evidence that neighborhood race has an effect on OHCA survival is limited. Other studies have examined the effect of neighborhood socioeconomic or deprivation indices and reported worse survival outcomes among patients with OHCAs in poorer neighborhoods.6, 7, 8, 9 These studies, however, used different composite socioeconomic indices that may not be readily accessible to policy makers unlike neighborhood income level, which can be readily obtained. Moreover, the likelihood of surviving an out‐of‐hospital cardiac arrest in a low‐, middle‐, or high‐income neighborhood independent of the racial composition of the neighborhood has also not been well characterized in prior studies but is important to understand to better define which types of communities cardiac arrest interventions should be targeted to reduce disparities in survival.
Accordingly, we leveraged data from CARES (the Cardiac Arrest Registry to Enhance Survival), which collects cardiac arrest data from a broad catchment area of more than 120 million US residents. We investigated differences in rates of survival to discharge and favorable neurological survival among predominantly white, integrated, and majority black neighborhoods as well as among low‐, middle‐, and high‐income neighborhoods. If survival differences by neighborhood race or income were present, we examined the extent to which survival differences were due to neighborhood differences in rates of bystander CPR and whether these differences were evident at the time of hospital admission. By better understanding these associations, we sought to determine to what extent out‐of‐hospital cardiac arrest survival differences by neighborhood race or income were largely due to care occurring outside the hospital.
Methods
The data that support the findings of this study are available from the corresponding author on request and approval by the CARES registry.
Data Sources and Study Design
CARES is a prospective, multicenter registry of patients with out‐of‐hospital cardiac arrest in the United States. Established by the Centers for Disease Control and Emory University for public health surveillance and continuous quality improvement, the design of the registry has been previously described.10, 11 Briefly, all patients with a confirmed out‐of‐hospital cardiac arrest (defined as pulselessness, apnea, and unresponsiveness) and for whom resuscitation is attempted are identified and followed from over 1400 EMS systems, representing a catchment area of ≈120 million US residents. Data are collected from 3 sources that, collectively, define the continuum of emergency cardiac care: 911 dispatch centers, EMS agencies, and receiving hospitals. Standardized international Utstein definitions for defining clinical variables and outcomes are used to ensure uniformity.12 A CARES analyst reviews every record for completeness and accuracy.11 The study was approved by Saint Luke's Mid America Heart Institute, which waived the requirement for informed consent because the analysis included only deidentified data.
Study Population
A total of 266 592 adults 18 years of age or older met criteria for an out‐of‐hospital cardiac arrest between January 1, 2013 and December 31, 2017 (Figure 1). We excluded 42 046 events that occurred in a residential healthcare facility because these sites typically have on‐site healthcare professionals. We further excluded 29 012 events that were witnessed by EMS personnel, 6592 events due to drowning, and 561 events with missing survival data. Additionally, we excluded 1732 events for which we were unable to link patients to a US census tract, yielding 186 649 out‐of‐hospital cardiac arrest events. We then excluded 17 147 events from 6875 census tracts with fewer than 5 total events during the study period. Our final cohort comprised 169 502 patients with an out‐of‐hospital cardiac arrest from 14 817 US census tracts.
Figure 1.
Definition of the study cohort. EMS indicates emergency medical services.
Data Collection and Processing
CARES collects patient‐level data on demographics (age, sex, and race), location of cardiac arrest, initial cardiac arrest rhythm, and whether the arrest was witnessed. Additionally, information is gathered as to whether bystander CPR was administered before EMS arrival and cardiac arrest etiology (presumed cardiac, respiratory, and other) is collected, as well as times to EMS arrival and duration of EMS treatment.
Survival outcomes were also collected by CARES. Our primary outcome was survival to hospital discharge. Secondary outcomes included survival to hospital admission and favorable neurological survival. Survival to hospital admission was defined as having return of spontaneous circulation prior to hospital arrival. Favorable neurological survival was defined as survival to hospital discharge with a cerebral performance category score of 1 (mild to no neurological disability) or 2 (moderate neurological disability).13 In addition, we examined rates of bystander CPR.
To obtain information on neighborhood race and income on each patient, CARES data were geocoded to a US census tract by the Centers for Disease Control on the basis of the address of the cardiac arrest. Census tracts are used as proxies for neighborhoods as they typically represent economically and socially homogeneous groups of ≈1200 to 8000 residents.14 Neighborhood variables were linked to each geocoded address with data from the 2010 US Census Summary Files and the 2016 American Community Survey's 5‐year estimates. From this linkage, the racial and income composition of each census tract neighborhood were identified. Categories of neighborhood race and income were defined a priori and were classified as predominantly white (>80% white), majority black (>50% black), or integrated. Integrated neighborhoods were those that met neither criterion for predominantly white or majority black. These cut points were chosen for interpretability but also mirrored the racial distribution in the study, as 19.3% of the neighborhoods were >50% black and 50% were >77.5% white. We likewise classified neighborhoods as low (median household income <$40 000), middle ($40 000 to $80 000), or high income (>$80 000), which also mirrored the distribution of the census tracts (lower quartile: ≤$40 716; middle 2 quartiles: $40 717 to $77 177; upper quartile: ≥$77 178). Additionally, we evaluated neighborhood race and income as continuous variables.
Statistical Analyses
The primary outcome was survival to hospital discharge. Baseline differences between those who survived to hospital discharge and those who died were compared using t tests for continuous variables and the χ2 test for categorical variables.
To assess that our categories for neighborhood race and income were robust, we first evaluated the relationship between neighborhood race and income with survival to discharge as continuous variables with a test of trend. Neighborhood race was categorized by deciles of proportion of black inhabitants, and neighborhood income was categorized by increments of $10 000 in median household family income. Then, to determine the association of neighborhood‐level race and income with survival, we constructed a hierarchical logistic regression model in which patient characteristics as well as neighborhood race and income were all modeled together as fixed effects, and the census tract neighborhood as a random effect, to address clustering of patients within neighborhoods. Neighborhood race (predominantly white, integrated, and majority black) and income (low, middle, and high) were categorized as previously outlined, and patient factors included patient age, sex, and race, whether the arrest was witnessed, location of cardiac arrest, and initial cardiac arrest rhythm. All covariates were retained in the model regardless of statistical significance. To determine whether differences in survival to discharge were mediated by differential rates of bystander CPR by neighborhood race, or income,4 we evaluated whether additional adjustment for bystander CPR attenuated survival differences. If survival differences remained, we explored whether this could be explained by differential EMS arrival times (time from activation to arrival of EMS) and treatment times (time from EMS arrival to departure) by neighborhood race and income by further including these variables in the model. Additionally, because rates of witnessed or shockable out‐of‐hospital cardiac arrest may differ by neighborhood, we repeated the above analyses in these patient subgroups.
Similar hierarchical regression models were constructed for the secondary end points of bystander CPR, survival to hospital admission, and favorable neurological survival at discharge. Discharge cerebral performance category scores were missing in 350 patients (0.2% of cohort). We used multiple imputation methods to impute missing values on the basis of all other observed data. Imputations were performed with Markov Chain Monte Carlo methods as implemented in SAS PROC MI (SAS Institute, Cary, NC). Ten imputed data sets were generated; analyses were replicated across data sets and pooled to obtain final estimates. Results with and without imputation were very similar; only the former are presented.
For each analysis, the null hypothesis was evaluated at a 2‐sided significance level of 0.05, and 95% CIs were calculated using robust standard errors. All statistical analyses were conducted using SAS Version 9.1.3 (SAS Institute, Cary, NC) and R Version 2.6.0 (Free Software Foundation, Boston, MA).
Results
Characteristics of the study cohort of 169 502 patients with out‐of‐hospital cardiac arrest are summarized in Table 1. The mean age of the study population was 62.8 years (SD 16.7 years), 63.3% were men, and 21.0% were of black race. Overall, 37.5% had had a cardiac arrest in a predominantly white neighborhood, 16.6% in a majority black neighborhood, and 45.9% in an integrated neighborhood, whereas 30.1%, 53.4%, and 16.5% had had cardiac arrests in low‐, middle‐, and high‐income neighborhoods, respectively. A total of 16 740 (9.9%) patients survived to hospital discharge. Patients of female sex and black race were less likely to survive to hospital discharge, whereas patients with a witnessed arrest, an arrest in a public location, bystander CPR, and an initial shockable cardiac arrest rhythm were more likely to survive to discharge.
Table 1.
Baseline Characteristics of Cohort, by Survival to Discharge
| Variables | Total | Survived to Discharge | Died | P Value |
|---|---|---|---|---|
| N=169 502 | n=16 740 | n=152 762 | ||
| Patient factors | ||||
| Age, y | <0.001 | |||
| Mean±SD | 62.8±16.7 | 58.0±15.4 | 63.3±16.7 | |
| Median (interquartile range) | 63.0 (52.0, 75.0) | 59.0 (49.0, 68.0) | 64.0 (53.0, 76.0) | |
| Sex | <0.001 | |||
| Male | 107 355 (63.3%) | 11 506 (68.7%) | 95 849 (62.7%) | |
| Female | 62 136 (36.7%) | 5234 (31.3%) | 56 902 (37.3%) | |
| Missing | 11 | 11 | ||
| Patient‐level race and ethnicity | <0.001 | |||
| White | 78 850 (46.5%) | 8398 (50.2%) | 70 452 (46.1%) | |
| Black | 35 632 (21.0%) | 2777 (16.6%) | 32 855 (21.5%) | |
| Asian | 3419 (2.0%) | 306 (1.8%) | 3113 (2.0%) | |
| Hispanic | 9580 (5.7%) | 850 (5.1%) | 8730 (5.7%) | |
| Native Hawaiian or Pacific Islander | 900 (0.5%) | 61 (0.4%) | 839 (0.5%) | |
| American Indian or Alaskan Native | 776 (0.5%) | 60 (0.4%) | 716 (0.5%) | |
| Unknown | 40 279 (23.8%) | 4287 (25.6%) | 35 992 (23.6%) | |
| Missing | 66 | 1 | 65 | |
| Location of cardiac arrest | <0.001 | |||
| Home | 139 937 (82.6%) | 10 361 (61.9%) | 129 576 (84.8%) | |
| Public or commercial building | 14 377 (8.5%) | 3336 (19.9%) | 11 041 (7.2%) | |
| Public street or highway | 10 473 (6.2%) | 1668 (10.0%) | 8805 (5.8%) | |
| Recreational facility | 2797 (1.7%) | 971 (5.8%) | 1826 (1.2%) | |
| Industrial place | 965 (0.6%) | 245 (1.5%) | 720 (0.5%) | |
| Other | 953 (0.6%) | 159 (0.9%) | 794 (0.5%) | |
| Was cardiac arrest witnessed | <0.001 | |||
| Bystander witnessed | 74 871 (44.2%) | 12 158 (72.6%) | 62 713 (41.1%) | |
| Unwitnessed | 94 631 (55.8%) | 4582 (27.4%) | 90 049 (58.9%) | |
| Cardiac arrest etiology | <0.001 | |||
| Presumed cardiac etiology | 150 791 (89.0%) | 14 487 (86.5%) | 136 304 (89.2%) | |
| Respiratory | 11 215 (6.6%) | 1310 (7.8%) | 9905 (6.5%) | |
| Other | 7496 (4.4%) | 943 (5.6%) | 6553 (4.3%) | |
| Person initiating CPR | <0.001 | |||
| First responder | 56 678 (33.4%) | 4363 (26.1%) | 52 315 (34.2%) | |
| Responding EMS personnel | 47 574 (28.1%) | 3363 (20.1%) | 44 211 (28.9%) | |
| Family member | 36 034 (21.3%) | 3703 (22.1%) | 32 331 (21.2%) | |
| Lay person | 23 665 (14.0%) | 4342 (25.9%) | 19 323 (12.6%) | |
| Lay medical provider | 5502 (3.2%) | 960 (5.7%) | 4542 (3.0%) | |
| Other | 45 (0.0%) | 9 (0.1%) | 36 (0.0%) | |
| Missing | 4 | 4 | ||
| Bystander CPR | 65 201 (38.5%) | 9005 (53.8%) | 56 196 (36.8%) | |
| First cardiac arrest rhythm | <0.001 | |||
| Nonshockable | ||||
| Asystole | 83 861 (49.5%) | 1690 (10.1%) | 82 171 (53.8%) | |
| PEA | 30 190 (17.8%) | 2530 (15.1%) | 27 660 (18.1%) | |
| Unknown nonshockable rhythm | 18 281 (10.8%) | 2120 (12.7%) | 16 161 (10.6%) | |
| Shockable | ||||
| Ventricular fibrillation | 26 697 (15.8%) | 7044 (42.1%) | 19 653 (12.9%) | |
| Ventricular tachycardia | 1545 (0.9%) | 457 (2.7%) | 1088 (0.7%) | |
| Unknown shockable rhythm | 8918 (5.3%) | 2895 (17.3%) | 6023 (3.9%) | |
| Missing | 10 | 4 | 6 | |
| Neighborhood factors | ||||
| Race of census tract | <0.001 | |||
| ≥80% White | 63 501 (37.5%) | 6749 (40.3%) | 56 752 (37.2%) | |
| ≥50% Black | 28 144 (16.6%) | 1946 (11.6%) | 26 198 (17.1%) | |
| Integrated | 77 857 (45.9%) | 8045 (48.1%) | 69 812 (45.7%) | |
| Median household income of census tract | <0.001 | |||
| <$40 000 annually | 51 087 (30.1%) | 4416 (26.4%) | 46 671 (30.6%) | |
| $40 000 to $80 000 annually | 90 480 (53.4%) | 9179 (54.8%) | 81 301 (53.2%) | |
| >$80 000 annually | 27 935 (16.5%) | 3145 (18.8%) | 24 790 (16.2%) | |
CPR indicates cardiopulmonary resuscitation; EMS, emergency medical services; PEA, pulseless electrical activity.
Study Outcomes
When assessed as continuous variables, both neighborhood race (P for trend <0.001) and neighborhood income (P for trend of 0.048) were each associated with the likelihood of a patient surviving to discharge after an out‐of‐hospital cardiac arrest (Figure 2). Unadjusted rates of survival to discharge by the study categories of neighborhood race and income are shown in Figure 3. Survival rates were similar for arrests in predominantly white and integrated neighborhoods but were lower in majority black neighborhoods. There was a gradient of lower survival in middle‐ and low‐income neighborhoods as compared with high‐income neighborhoods. After multivariable adjustment, patients with an out‐of‐hospital cardiac arrest in a majority black neighborhood were 12% less likely (adjusted odds ratio [OR] 0.88; 95% CI 0.82‐0.95; P<0.001) to survive to discharge as compared with those in a predominantly white neighborhood, whereas there was no difference in survival between integrated and predominantly white neighborhoods (Table 2). As compared with high‐income neighborhoods, those in middle‐income neighborhoods were 11% (adjusted OR 0.89; 95% CI 0.85‐0.94; P<0.001) less likely to survive to discharge, whereas those in low‐income neighborhoods were 12% (adjusted OR 0.88; 95% CI 0.83‐0.94; P<0.001) less likely to survive. Adjusted estimates for different combinations of neighborhood race and income strata are provided in Table 3. Notably, lower rates of survival to discharge in black patients (no adjustment for neighborhood factors: adjusted OR for black versus white patients 0.86; 95% CI 0.82‐0.91; P>0.001) were markedly attenuated after adjustment for neighborhood race and income (adjusted OR for black versus white patients 0.95; 95% CI 0.89‐1.02; P=0.19). The full model is summarized in Table S1.
Figure 2.
Neighborhood race and income for survival to discharge. Tests of trend of the relationship between neighborhood race (A) and income (B) and survival to discharge are shown. Numbers above each vertical bar represent number of US census tracts. OHCA indicates out‐of‐hospital cardiac arrest.
Figure 3.
Unadjusted rates of survival outcomes by categories of neighborhood race and income.
Table 2.
Association Between Neighborhood Race and Income and Survival Outcomes
| Outcome | Neighborhood Race | ||||||
|---|---|---|---|---|---|---|---|
| >80% Whitea (n=63 501) | >50% Black (n=28 144) | Adjusted OR (95% CI) | P Value | Integrated (n=77 857) | Adjusted OR (95% CI) | P Value | |
| Primary outcome | |||||||
| Survival to hospital discharge | 6749 (10.6%) | 1946 (6.9%) | 0.88 (0.82, 0.95) | <0.001 | 8045 (10.3%) | 1.00 (0.96, 1.04) | 0.93 |
| Adjusted for bystander CPR | 0.91 (0.84, 0.98) | 0.009 | 1.01 (0.97, 1.05) | 0.77 | |||
| Secondary outcomes | |||||||
| Bystander CPR | 27 764 (43.7%) | 6962 (24.7%) | 0.60 (0.57, 0.63) | <0.001 | 30 475 (39.1%) | 0.88 (0.85, 0.91) | <0.001 |
| Survival to hospital admission | 17 470 (27.5%) | 6478 (23.0%) | 0.98 (0.93, 1.03) | 0.34 | 22 603 (29.0%) | 1.11 (1.08 1.14) | <0.001 |
| Favorable neurological discharge | 5792 (9.1%) | 1203 (4.3%) | 0.76 (0.70, 0.83) | <0.001 | 6457 (8.3%) | 0.97 (0.92, 1.01) | 0.14 |
| Adjusted for bystander CPR | 0.79 (0.72, 0.80) | <0.001 | 0.97 (0.93, 1.02) | 0.25 | |||
| Neighborhood Income | |||||||
|---|---|---|---|---|---|---|---|
| >$80 000a (n=27 935) | <$40 000 (n=51 087) | Adjusted OR (95% CI) | P Value | $40 000 to $80 000 (n=90 480) | Adjusted OR (95% CI) | P Value | |
| Primary outcome | |||||||
| Survival to hospital discharge | 3145 (11.3%) | 4416 (8.6%) | 0.88 (0.83, 0.94) | <0.001 | 9179 (10.1%) | 0.89 (0.85, 0.94) | <0.001 |
| Adjusted for bystander CPR | 0.90 (0.85, 0.96) | 0.001 | 0.90 (0.86, 0.95) | <0.001 | |||
| Secondary outcomes | |||||||
| Bystander CPR | 12 448 (44.6%) | 15 496 (30.3%) | 0.67 (0.64, 0.70) | <0.001 | 37 257 (41.2%) | 0.88 (0.85, 0.92) | <0.001 |
| Survival to hospital admission | 7975 (28.5%) | 13 783 (27.0%) | 1.03 (0.99, 1.07) | 0.17 | 24 793 (27.4%) | 0.94 (0.91, 0.97) | <0.001 |
| Favorable neurological discharge | 2684 (9.6%) | 3136 (6.2%) | 0.80 (0.75, 0.85) | <0.001 | 7632 (8.4%) | 0.87 (0.82, 0.92) | <0.001 |
| Adjusted for bystander CPR | 0.82 (0.77, 0.88) | <0.001 | 0.88 (0.83, 0.93) | <0.001 | |||
All models adjusted for fixed effects of patient age, sex, and race, whether the arrest was witnessed, location of cardiac arrest, and initial cardiac arrest rhythm as well as neighborhood race and neighborhood income. Census tract neighborhood was modeled as random effect in the hierarchical models. CPR indicates cardiopulmonary resuscitation; OR, odds ratio.
Reference groups for the neighborhood race and income analyses are predominantly (>80%) white and high‐income (median household income >$80 000) neighborhoods.
Table 3.
Likelihood of Survival to Discharge by Different Combinations of Neighborhood Race and Income Strata
| Neighborhood Race and Income Stratum | Adjusted OR (95% CI) | P Value |
|---|---|---|
| >80% White and >$80 000 | Reference | Reference |
| Low income (<$40 000) | ||
| >50% Black | 0.81 (0.74, 0.89) | <0.001 |
| Integrated | 0.84 (0.78, 0.91) | <0.001 |
| >80% White | 0.89 (0.80, 0.99) | 0.03 |
| Middle income ($40 000 to $80 000) | ||
| >50% Black | 0.73 (0.65, 0.83) | <0.001 |
| Integrated | 0.90 (0.84, 0.96) | 0.002 |
| >80% White | 0.84 (0.79, 0.90) | <0.001 |
| High income (>$80 000) | ||
| >50% black | NAa | NAa |
| Integrated | 0.92 (0.84, 1.00) | 0.058 |
OR indicates odds ratio.
No estimate provided as there were only 533 patients (0.3% of entire cohort) in this stratum, making model estimates for this group unreliable.
Rates of bystander CPR differed markedly by neighborhood factors. After adjustment for patient and cardiac arrest characteristics, patients in majority black and integrated neighborhoods were respectively 40% and 12% less likely to have CPR initiated by a bystander as compared with patients in predominantly white neighborhoods (see Table 2). Similarly, those in low‐ and middle‐income neighborhoods were 33% and 12% less likely to receive bystander CPR as compared with those in high‐income neighborhoods. However, inclusion of bystander CPR as a patient‐level variable in the model for survival to discharge did not substantially attenuate survival differences between predominantly white and majority black neighborhoods nor between high‐ and middle‐ or low‐income neighborhoods (see Table 2).
There were small differences by neighborhood race and income in EMS arrival times. In contrast, EMS treatment times were longer in predominantly white neighborhoods (median of 23.4, 19.7, and 21.0 minutes [P<0.001] in predominantly white, majority black, and integrated neighborhoods, respectively) and high‐income neighborhoods (median of 22.8, 22.0, and 21.0 minutes [P<0.001] in high‐, middle‐, and low‐income neighborhoods, respectively) (Table S2). Nonetheless, additional adjustment for EMS arrival and treatment times among those patients with complete data on EMS times did not significantly attenuate survival differences by neighborhood race or income (Table S3).
Rates of favorable neurological discharge mirrored those for survival to discharge (see Figure 3 and Table 2). Patients with an out‐of‐hospital cardiac arrest in a majority black neighborhood were 24% less likely to survive without severe neurological disability as compared with predominantly white communities, whereas there was no difference between integrated and predominantly white neighborhoods. Rates of favorable neurological discharge were also lower in middle‐ and low‐income neighborhoods as compared with high‐income neighborhoods. As with the outcome of survival to discharge, these differences by neighborhood race and income were only modestly attenuated after neighborhood differences in rates of bystander CPR had been accounted for (see Table 2).
Despite differences by neighborhood race and income in overall survival and favorable neurological discharge, rates of survival to hospital admission were not different between out‐of‐hospital cardiac arrest patients from predominantly white and majority black neighborhoods and from high‐income and low‐income neighborhoods. Moreover, there were only small differences in rates of survival to hospital admission between high‐income and medium‐income neighborhoods, and rates of survival to hospital admission were actually higher in integrated neighborhoods as compared with predominantly white neighborhoods (see Table 2). Finally, all study findings were similar when the analyses were restricted to the 74 871 (44.2%) out‐of‐hospital cardiac arrests that were witnessed by a bystander or to the 37 160 (21.9%) patients with an initial shockable cardiac arrest rhythm (Tables S4 and S5).
Discussion
We found that the racial composition and median income of a neighborhood influences the likelihood that an individual with an out‐of‐hospital cardiac arrest survives to hospital discharge or has favorable neurological survival. Patients with an out‐of‐hospital cardiac arrest in a majority black neighborhood were 12% less likely to survive to discharge and 24% less likely to survive without severe neurological disability than those in a predominantly white neighborhood, whereas there were no differences in either outcome between integrated and predominantly white neighborhoods. In contrast, rates of survival to discharge and favorable neurological survival were both lower in low‐ and middle‐income neighborhoods as compared with high‐income neighborhoods. Importantly, similar patterns were found in the population with a witnessed out‐of‐hospital cardiac arrest, suggesting that survival differences were not simply due to differences in witnessed cardiac arrest events by neighborhood race or income, as witnessed events typically are the opportunities to provide bystander CPR.
Few studies have evaluated the effect of neighborhood race on out‐of‐hospital cardiac arrest outcomes. One early study in CARES found lower rates of bystander CPR in communities that were largely black or low income but did not examine survival, as it was likely underpowered given a sample size of 14 225 patients in that study (or <9% of the current cohort).4 A more recent study of 22 816 out‐of‐hospital cardiac arrests from the Resuscitation Outcome Consortium reported much lower odds of survival to discharge in largely black communities than were found in this study (ORs of 0.65 and 0.68 for communities that were >75% black and 51% to 75% black, as compared with those that were <25% black), but the larger relative differences may have been due to the limited geographical regions examined in that study.5 Our findings extend the findings of these 2 prior studies in a much larger cohort of patients with out‐of‐hospital cardiac arrest and in a catchment area representing approximately one third of the United States. We confirmed differences in survSival in largely black neighborhoods but found no survival differences in patients from integrated neighborhoods. Moreover, we provided explicit reporting on the effect of the median income of a neighborhood on several survival outcomes for out‐of‐hospital cardiac arrest, which has been reported largely for the outcome of receipt of bystander CPR.15 Although studies on the effect of composite neighborhood socioeconomic indices on OHCA outcomes exist,6, 7, 8, 9 these composite indices may not be as readily accessible as neighborhood income.
Our findings suggest important deficiencies in our current public health approach in responding to out‐of‐hospital cardiac arrest. Low survival rates for this condition are unsurprising, but potential disparities in survival based solely on the racial and income composition of the neighborhood in which one has a cardiac arrest raise concerns. Although differences in CPR training16 and bystander CPR delivery4, 5 have been reported by neighborhood race and income, we found that these differences accounted for only a small fraction of the overall survival differences between majority black and predominantly white neighborhoods and between low‐ and middle‐income versus high‐income neighborhoods. Other reasons for survival differences by neighborhood may be related to differences in EMS treatment, and we found substantially longer EMS treatment times in predominantly white and high‐income neighborhoods than in their counterpart comparisons. However, additional adjustment for EMS arrival and treatment times did not attenuate survival differences. This suggests that other aspects of resuscitation care may differ by neighborhood, including paramedic training, number of EMS responders, CPR quality, and delivery of other aspects of acute cardiac life support. These resuscitation care factors were not collected within CARES and are difficult to quantify but certainly deserve closer scrutiny, as it is unclear from our findings if a sole focus of merely increasing CPR training and bystander CPR delivery in majority black and low‐ and middle‐income neighborhoods is sufficient to eliminate the differences in survival outcomes by neighborhood race and income.
On the other hand, our findings that rates of survival to hospital admission did not differ by neighborhood race and yielded only small differences by neighborhood income may also suggest that overall survival differences may be due to neighborhood differences in postresuscitation hospital care.17 Indeed, prior studies have reported that racial differences in in‐hospital cardiac arrest survival are, in part, attributable to the racial composition of the hospital at which one receives care,18 although the reasons for this (eg, differences in resources or expertise in intensive care) are unclear. Therefore, efforts to reduce neighborhood disparities in survival for out‐of‐hospital cardiac arrest may also need to address differences in hospital care. Of course, some of the difference in rates of overall survival and favorable neurological survival may still be due to prehospital care, especially if the clinical impact of neighborhood differences in EMS treatment times, bystander CPR, CPR quality, and other aspects of resuscitation care are not fully appreciated until after hospital arrival.
Finally, it is notable that lower rates of survival to discharge for black patients with out‐of‐hospital cardiac arrest were largely attenuated after controlling for the racial and income composition of the neighborhood in which the individual had the event (OR for black versus white patients went from 0.86 to 0.95). This indicates that a substantial proportion of the existing racial disparities in survival for out‐of‐hospital cardiac arrest may be modifiable if effective interventions are designed and implemented in the most vulnerable communities. To date, however, most efforts, such as CPR training, have been performed more often in white and wealthier communities.16 This suggests a critical need to develop systems of care to deliver training and interventions to those communities with the lowest cardiac arrest survival rates, many of which are predominantly black and non–high income.
Our study should be interpreted in the context of the following limitations. First, our study was an observational study that evaluated associations between neighborhood race and income and survival outcomes. These associations may be due to neighborhood differences in resources such as EMS funding and basic life support training for bystander CPR, which may be the focus of future public health policy initiatives to address neighborhood disparities in survival outcomes for out‐of‐hospital cardiac arrest. However, they may also be due, in part, to neighborhood‐level differences in comorbidities. These are not collected within CARES to minimize the burden of data submission given the large volume of out‐of‐hospital cardiac arrest cases collected annually and may represent unmeasured confounding. Second, although we were able to define survival differences by neighborhood race and income and assess the impact of neighborhood differences in bystander CPR rates, we were unable to fully account for the differences in survival because information on many aspects of acute resuscitation care before hospital arrival was not available. Moreover, data on EMS arrival and treatment times were incomplete. Although we conducted sensitivity analyses to examine their impact, these sensitivity analyses should be interpreted with some caution. Third, CARES does not systematically collect data on postresuscitation intensive care as it is primarily an out‐of‐hospital cardiac arrest registry. Although our findings suggest that some of the survival differences by neighborhood race and income were likely due to postresuscitation care in hospitals, we were not able to identify which aspects of postresuscitation care contributed to such survival differences. Finally, although this study encompassed communities representing nearly one third of the US population, our findings may not apply to communities that do not participate in CARES.
In conclusion, we found that a neighborhood's racial and income composition influences the likelihood that an individual survives an out‐of‐hospital cardiac arrest. Bystander CPR accounted for some but not all of these differences, and unmeasured patient illness severity by race and income could partially explain our findings. Nonetheless, our findings suggest that differences in other aspects of EMS response that differ by neighborhood race and income may be important to address. Moreover, the lack of neighborhood differences in rates of survival to hospital admission implies that overall survival differences may also be due to differences in postresuscitation care at hospitals. Our findings suggest the need to improve not only bystander CPR rates but also other aspects of EMS treatment and postresuscitation care in communities that are majority black and non–high income.
Sources of Funding
Dr Chan is supported by an R01 grant (1R01HL123980) from the National Heart, Lung, and Blood Institute.
Disclosures
Dr Chan has received funding support from the American Heart Association, which currently helps to fund the CARES registry. CARES was funded by the Centers for Disease Control and Prevention from 2004 to 2012. The program is now supported through private funding from the American Red Cross, the American Heart Association, and in‐kind support from Stryker Physio‐Control and Emory University. Dr McNally is supported by grant funding from CARES and serves as Executive Director of the program. None of these funding partners had a role in the study design, data analysis, or manuscript preparation and revision. The remaining authors have no disclosures to report.
Supporting information
Table S1. Detailed Model With Patient Factors and Neighborhood Race and Income
Table S2. EMS Arrival and Treatment Times by Neighborhood Race and Median Household Income
Table S3. Sequential Models for the Association Between Neighborhood Race and Income and Survival to Discharge
Table S4. Survival Outcomes by Neighborhood Race and Income for Witnessed Out‐of‐Hospital Cardiac Arrests
Table S5. Survival Outcomes by Neighborhood Race and Income for Shockable Out‐of‐Hospital Cardiac Arrests
Acknowledgments
Dr Chan had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
(J Am Heart Assoc. 2020;9:e014178 DOI: 10.1161/JAHA.119.014178.)
References
- 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 Committee and Stroke Statistics Subcommittee . Heart disease and stroke statistics—2012 update: a report from the American Heart Association. Circulation. 2012;125:e2–e220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Chan PS, McNally B, Tang F, Kellermann A. Recent trends in survival from out‐of‐hospital cardiac arrest in the United States. Circulation. 2014;130:1876–1882. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. 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. Regional variation in out‐of‐hospital cardiac arrest incidence and outcome. JAMA. 2008;300:1423–1431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Sasson C, Magid DJ, Chan P, Root ED, McNally BF, Kellermann AL, Haukoos JS. Association of neighborhood characteristics with bystander‐initiated CPR. N Engl J Med. 2012;367:1607–1615. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Starks MA, Schmicker RH, Peterson ED, May S, Buick JE, Kudenchuk PJ, Drennan IR, Herren H, Jasti J, Sayre M, Stub D, Vilke GM, Stephens SW, Chang AM, Nuttall J, Nichol G; Resuscitation Outcomes Consortium . Association of neighborhood demographics with out‐of‐hospital cardiac arrest treatment and outcomes: where you live may matter. JAMA Cardiol. 2017;2:1110–1118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Ahn KO, Shin SD, Hwang SS, Oh J, Kawachi I, Kim YT, Kong KA, Hong SO. Association between deprivation status at community level and outcomes from out‐of‐hospital cardiac arrest: a nationwide observational study. Resuscitation. 2011;82:270–276. [DOI] [PubMed] [Google Scholar]
- 7. Buick JE, Ray JG, Kiss A, Morrison LJ. The association between neighborhood effects and out‐of‐hospital cardiac arrest outcomes. Resuscitation. 2016;103:14–19. [DOI] [PubMed] [Google Scholar]
- 8. 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;79:417–423. [DOI] [PubMed] [Google Scholar]
- 9. Lee SY, Song KJ, Shin SD, Ro YS, Hong KJ, Kim YT, Hong SO, Park JH, Lee SC. A disparity in outcomes of out‐of‐hospital cardiac arrest by community socioeconomic status: a ten‐year observational study. Resuscitation. 2018;126:130–136. [DOI] [PubMed] [Google Scholar]
- 10. McNally B, Stokes A, Crouch A, Kellermann AL. CARES: cardiac arrest registry to enhance survival. Ann Emerg Med. 2009;54:674–683.e672. [DOI] [PubMed] [Google Scholar]
- 11. McNally B, Robb R, Mehta M, Vellano K, Valderrama AL, Yoon PW, Sasson C, Crouch A, Perez AB, Merritt R, Kellermann A. Out‐of‐hospital cardiac arrest surveillance—Cardiac Arrest Registry to Enhance Survival (CARES), United States, October 1, 2005–December 31, 2010. MMWR Surveill Summ. 2011;60:1–19. [PubMed] [Google Scholar]
- 12. Jacobs I, Nadkarni V, Bahr J, Berg RA, Billi JE, Bossaert L, Cassan P, Coovadia A, D'Este K, Finn J, Halperin H, Handley A, Herlitz J, Hickey R, Idris A, Kloeck W, Larkin GL, Mancini ME, Mason P, Mears G, Monsieurs K, Montgomery W, Morley P, Nichol G, Nolan J, Okada K, Perlman J, Shuster M, Steen PA, Sterz F, Tibballs J, Timerman S, Truitt T, Zideman D. Cardiac arrest and cardiopulmonary resuscitation outcome reports: update and simplification of the Utstein templates for resuscitation registries: a statement for healthcare professionals from a Task Force of the International Liaison Committee on Resuscitation (American Heart Association, European Resuscitation Council, Australian Resuscitation Council, New Zealand Resuscitation Council, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Councils of Southern Africa). Circulation. 2004;110:3385–3397. [DOI] [PubMed] [Google Scholar]
- 13. Jennett B, Bond M. Assessment of outcome after severe brain damage. Lancet. 1975;1:480–484. [DOI] [PubMed] [Google Scholar]
- 14. United States Census Bureau . Geographic terms and concepts. Available at: https://www.census.gov/geo/reference/gtc/gtc_place.html. Accessed September 5, 2018.
- 15. Sasson C, Keirns CC, Smith DM, Sayre MR, Macy ML, Meurer WJ, McNally BF, Kellermann AL, Iwashyna TJ. Examining the contextual effects of neighborhood on out‐of‐hospital cardiac arrest and the provision of bystander cardiopulmonary resuscitation. Resuscitation. 2011;82:674–679. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Anderson ML, Cox M, Al‐Khatib SM, Nichol G, Thomas KL, Chan PS, Saha‐Chaudhuri P, Fosbol EL, Eigel B, Clendenen B, Peterson ED. Rates of cardiopulmonary resuscitation training in the United States. JAMA Intern Med. 2014;174:194–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Meaney PA, Bobrow BJ, Mancini ME, Christenson J, de Caen AR, Bhanji F, Abella BS, Kleinman ME, Edelson DP, Berg RA, Aufderheide TP, Menon V, Leary M. Cardiopulmonary resuscitation quality: [corrected] improving cardiac resuscitation outcomes both inside and outside the hospital: a consensus statement from the American Heart Association. Circulation. 2013;128:417–435. [DOI] [PubMed] [Google Scholar]
- 18. Chan PS, Nichol G, Krumholz HM, Spertus JA, Jones PG, Peterson ED, Rathore SS, Nallamothu BK. Racial differences in survival after in‐hospital cardiac arrest. JAMA. 2009;302:1195–1201. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Detailed Model With Patient Factors and Neighborhood Race and Income
Table S2. EMS Arrival and Treatment Times by Neighborhood Race and Median Household Income
Table S3. Sequential Models for the Association Between Neighborhood Race and Income and Survival to Discharge
Table S4. Survival Outcomes by Neighborhood Race and Income for Witnessed Out‐of‐Hospital Cardiac Arrests
Table S5. Survival Outcomes by Neighborhood Race and Income for Shockable Out‐of‐Hospital Cardiac Arrests