C

Ao

KYa

b

c

d

e

f

a

ARRA

KCSOD

1

oicthI

i

N

(ij

0d

Resuscitation 82 (2011) 270–276

Contents lists available at ScienceDirect

Resuscitation

journa l homepage: www.e lsev ier .com/ locate / resusc i ta t ion

linical paper

ssociation between deprivation status at community level and outcomes fromut-of-hospital cardiac arrest: A nationwide observational study�,��

i Ok Ahnc, Sang Do Shina,b,∗, Seung Sik Hwangd, Juhwan Ohb,e, Ichiro Kawachie,oung Taek Kimf, Kyoung Ae Kongf, Sung Ok Hongf

Department of Emergency Medicine, Seoul National University College of Medicine, 101 Daehakro, Jongno-Gu, Seoul 110-744, Republic of KoreaInstitute of Health Policy and Management, Medical Research Center, College of Medicine, Seoul National University, Seoul, Republic of KoreaCenter for Education and Training of EMS and Rescue, Seoul Fire Academy, Seoul, Republic of KoreaDepartment of Preventive Medicine, Inha University College of Medicine, Republic of KoreaDepartment of Society, Human Development, and Health, Harvard School of Public Health, Boston, MA, USAKorea Centers for Disease Control and Prevention, Republic of Korea

r t i c l e i n f o

rticle history:eceived 27 July 2010eceived in revised form 19 October 2010ccepted 28 October 2010

eywords:ardiac arrestocioeconomic statusutcomeseprivation index

a b s t r a c t

Study objectives: We sought to examine the association between area deprivation and outcomes of out-of-hospital cardiac arrest in Korea.Methods: Data were obtained from the emergency medical service (EMS) system. A nationwide OHCAcohort database from January2006 to December 2007 was constructed via hospital chart review andambulance run sheet data. We enrolled all EMS-assessed OHCA victims and excluded cases withoutavailable hospital outcome data or residential address. The Carstairs index was used to categorize dis-tricts according to level of deprivation into five quintiles, from (Q1, the least deprived) to (Q5, themost deprived). Main outcomes were survival to hospital discharge, survival to admission, and return ofspontaneous circulation (ROSC).Results: 34,227 patients were included. Initial rhythm, witnessed status, attempted bystander cardiopul-

monary resuscitation (CPR), CPR by EMS, CPR in the emergency department (ED), and elapsed timeinterval significantly varied according to area deprivation level (p < 0.001). OHCA outcomes were consis-tently worse in the most deprived areas. The adjusted OR (95% CI) for survival to hospital discharge was0.58 (0.45–0.77) in the most deprived areas compared to the least deprived areas.

epriv

Conclusion: Community darrest patients in Korea.

. Introduction

An estimated 166,000–310,000 out-of-hospital cardiac arrestsccur in the United States each year.1 A set of conditions includ-ng being witnessed at the time of the cardiac arrest, attempted

ardiopulmonary resuscitation (CPR) by bystanders, shockable ini-ial rhythm, and shorter time interval elapsed between arrest andospitalization are established predictors for better outcomes.2

mproving survival from out-of-hospital cardiac arrest to 16.3%

� A Spanish translated version of the summary of this article appears as Appendixn the final online version at doi:10.1016/j.resuscitation.2010.10.023.�� Previous presentation: This paper was presented at the annual meeting of theational Association of EMS Physicians in Phoenix, AZ in 2010 January.∗ Corresponding author. Tel.: +82 2 2072 0854; fax: +82 2 741 7855.

E-mail addresses: arendt75@gmail.com (K.O. Ahn), shinsangdo@medimail.co.krS.D. Shin), cyberdoc@inha.ac.kr (S.S. Hwang), oh328@snu.ac.kr (J. Oh),kawachi@hsph.harvard.edu (I. Kawachi), ruyoung@cdc.go.kr (Y.T. Kim),oy-well@hanmail.net (K.A. Kong), soh822@hanmail.net (S.O. Hong).

300-9572/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.resuscitation.2010.10.023

ation was strongly associated with survival among out-of-hospital cardiac

© 2010 Elsevier Ireland Ltd. All rights reserved.

throughout North America (the best observed rate) would avertan estimated 15,000 premature deaths annually.3

Socioeconomic status (SES) is a robust predictor of outcomesin cardiovascular disease4,5 although the association between SESand out-of-hospital cardiac arrest (OHCA) has not been consistentacross studies.6–8 Some studies reported no association betweenmedian income or race and outcomes from OHCA7 while othershave suggested a racial disparity for incidence and outcome ofOHCA.8,9 Existing studies have mainly focused on individual-levelSES (as well as race/ethnic status) as a predictor for OHCA outcomes.With few exceptions, community-level SES has not been investi-gated as a predictor of outcomes following OHCA.6,10 They foundthat individuals suffering an OHCA in a predominantly AfricanAmerican community were less likely to have CPR attempted by

a bystander. There are several potential mechanisms to account forarea variations in the outcomes of OHCA, including: (a) residentsin communities with lower levels of health literacy may not knowhow to perform CPR, (b) residents in communities where OHCAis more common are better prepared to act quickly (to the extent

citatio

tbbgihwn

cmn

2

2

2ahgvfiac(taarpaaiTg

mrslwlpe((n

maaegta

it(iip

K.O. Ahn et al. / Resus

hat cardiac arrest is more common in more disadvantaged neigh-orhoods, this mechanism would predict that OHCA outcomes areetter in more deprived areas), (c) lower accessibility to emer-ency care, or quality care, may result in worse OHCA prognosisn resource-poor communities (to the extent that major teachingospitals are located in low SES neighborhoods, this mechanismould also predict that OHCA outcomes would be better in sucheighborhoods).

The present study examined whether community SES is asso-iated with OHCA outcomes in South Korea, using an emergencyedical service (EMS) database that recorded all cases in the

ational population during 2006 and 2007.

. Methods

.1. Study setting

South Korea, with a total population of 48.6 million (in005), experiences about 1.4 million calls for ambulance servicesnnually, of which an estimated 20,000 (in 2007) are for out-of-ospital cardiac arrest. The Korean EMS system is single-tier andovernment-provided, and basic life support (BLS) ambulance ser-ices are operated by 16 provincial headquarters of the nationalre department. Ambulance crews are trained to administer CPRt the scene and during transport, and in limited areas can provideare comparable to intermediate emergency medical techniciansEMT-I) in the US. This includes inserting intravenous lines, endo-raheal intubation, classic type laryngeal mask airway insertion,nd the administration of medications including epinephrine andtropine under direct medical supervision. Equipment and mate-ials are standardized under the Korean EMS Act. However, suchre-hospital advanced life support is not uniformly available acrossll areas of the country; in most areas, advanced life support is onlyvailable in hospitals. The EMTs cannot declare death or stop CPRn the field unless return of spontaneous circulation (ROSC) occurs.herefore, all patients with OHCA are transported to hospital emer-ency departments (EDs).11

All EDs are formally designated as level 1–3 by the govern-ent. The designation is mostly based on the ED’s level of human

esources, essential instruments and equipment, and service levelsuch as availability of certain specialists. Wilderness areas and iso-ated islands usually have no designated ED and here most patients

ith OHCA are transported to the regional health care centers. Mostevel 3 EDs are not well equipped and are usually served by generalhysicians. However, level 1 and level 2 EDs must be covered bymergency physicians 24 h a day by law. There are 20 regional EDslevel 1), 99 local EDs (level 2), about 300 small emergency roomsERs) (level 3), and about 400 non-ED facilities that still treat smallumbers of emergency patients.12

The national EMS plan was initiated by the national govern-ent through a specific EMS fund which was founded in 2002

nd expanded with funding up to $2 billion (in US dollars equiv-lent) in 2009. This fund was used for supporting designatedmergency departments, EMS development, quality assurance pro-rams, and research programs. The EMS fund was also targetedo improve the structure and process of EMS in selected ruralreas.

The Korean National Health System subjects were classifiednto two categories, according to whether they are covered by

he National Health Insurance (NHI) (96.85%) or by Medicaid (MA)3.15%). The NHI with MA covers the entire Korean populationncluding OHCA. The NHI covers about 70% of medical cost dur-ng hospital care for OHCA while remained 25% should by paid byatients themselves.

n 82 (2011) 270–276 271

2.2. Data collection

We used a nationwide, population-based, and EMS-assessedOHCA database (= CAVAS database) with coverage of the entirecountry.11 The CAVAS database was built from ambulance runsheets and followed by hospital record review. The databaseconstruction began in 2006 and continues to the present daywith support from the Korean Centers for Disease Control andthe National Emergency Management Agency. The ambulance-run sheet is electronically recorded whenever an ambulance isdispatched, and the recorded data are stored in the province head-quarters data server of the fire department. It contains geographicaland socio-demographic data, place of cardiac arrest, elapsed timevariables associated with resuscitation efforts, content of treat-ments, and the destination hospitals. The review of hospital recordswas conducted by a Korean CDC expert trained in medical recordreview. To maintain the quality of medical review in instances ofOHCA, a quality maintenance committee was established consist-ing of emergency physicians, statistical experts, epidemiologists,and medical record reviewers. Designated education and trainingfor this study were provided for all reviewers over three days, andscenario-based pilot reviews were developed and conducted by thereviewers. The review form followed the Utstein Style report formand was customized to this study setting.2

The details on data collection process have been previously.11

This study was approved by the Institutional Review Board of thestudy hospital, Seoul National University Hospital.

2.3. Selection and description of participants

We enrolled all cases with OHCA who were transported to alllevels of EDs (level 1 to level 3 and non-Ed facilities) from January1, 2006 to December 31, 2007 from the CAVAS database. Vari-ables extracted from the dataset included general demographics,causes of arrest, whether bystander CPR was attempted, initial ECGrhythm, witnessed status, whether CPR was performed by the EMSor at the ED, elapsed time intervals, and geographic location wherethe OHCA occurred. When there was no information in the medicalrecord review, we assigned these cases to “unknown” status. Weanalyzed the data including these unknown groups, not omittingthe cases. Elapsed response time intervals were calculated as thetime from the initial call to the arrival of ambulance at the scene,and total transport time intervals from the initial call to ED arrival.

Geographic information was obtained from the EMS agencyoperating ambulance service. All cases were coded with districtlevel information. Average population number per district was219,542 (standard deviation: 187,949) people.

2.4. Socioeconomic status

For community-level SES measures, area deprivation indicesbased on the Carstairs Index13 were calculated for each of the 250districts across the country. On the basis of the National Census Datain 2005, we calculated the district-specific proportions of house-holds with: (1) overcrowding (more than 1.5 persons/room); (2) thepercent unemployment among economically active men (between15 and 64 years); (3) the percent in manual occupations; and (4)lack of car ownership. We finally calculated the district-specificdeprivation index by averaging the z-standardized scores acrossthe Census-derived four indicators. The districts were then cate-

gorized into five quintiles, based on the level of deprivation, fromQ1 (least deprived) to Q5 (most deprived). Every district was fur-ther classified according to whether they were metropolitan, urban,or rural, using guidelines set by the Ministry of Public Affairs andSecurity (Table 1).

272 K.O. Ahn et al. / Resuscitatio

Tab

le1

Nu

mbe

rof

dis

tric

tan

dp

opu

lati

onn

um

ber

acco

rdin

gto

dep

riva

tion

ind

excl

assi

fica

tion

in20

06.

Urb

aniz

atio

nTo

tal

Q1

(low

est)

Q2

(low

)Q

3(m

idd

le)

Q4

(hig

h)

Q5

(hig

hes

t)

Dis

tric

t,N

Pop

ula

tion

,ND

istr

ict,

NPo

pu

lati

on,N

Dis

tric

t,N

Pop

ula

tion

,ND

istr

ict,

NPo

pu

lati

on,N

Dis

tric

t,N

Pop

ula

tion

,ND

istr

ict,

NPo

pu

lati

on,N

Met

rop

olit

an95

29,0

76,1

6220

7,05

9,43

421

7,41

0,50

221

6,84

7,31

419

5,44

6,25

414

2,31

2,65

8U

rban

6713

,504

,502

122,

240,

363

133,

164,

795

153,

572,

657

101,

765,

979

172,

760,

708

Ru

ral

884,

460,

770

1897

5,49

416

788,

479

1466

6,14

021

981,

876

191,

048,

781

Tota

l25

047

,041

,434

5010

,275

,291

5011

,363

,776

5011

,086

,111

508,

194,

109

506,

122,

147

OH

CA

,tot

alfo

rtw

oye

ar34

,227

6499

7868

7650

6490

5720

Cru

de

inci

den

cera

tep

eron

eye

ar

36.4

31.6

34.6

34.5

39.6

46.7

aM

etro

pol

itan

,urb

an,a

nd

rura

lwas

defi

ned

onth

eba

sis

ofcl

assi

fica

tion

ofK

orea

Nat

ion

alSt

atis

tica

lOffi

ce.M

etro

pol

itan

city

has

the

pop

ula

tion

nu

mbe

rov

erth

an1.

0m

illi

on.U

rban

city

has

50,0

00to

1,00

0,00

0p

opu

lati

onn

um

ber.

Ru

ralc

oun

tyh

asle

ssth

an50

,000

pop

ula

tion

nu

mbe

r.

n 82 (2011) 270–276

2.5. Outcome measure

The primary outcome variable of interest was survival to hospi-tal discharge. Survival to discharge was defined when a patient wasdischarged from hospital or transferred to other facilities includ-ing rehabilitation centers. We additionally examined survival tohospital admission and return of spontaneous circulation (ROSC).

2.6. Statistical analysis

Outcomes according to the area deprivation level were com-pared using the Cochrane–Armitage test for categorical variablesfor trend and the ANOVA test with post hoc analysis for continuousvariables. Multivariate logistic regression adjusting for covariateswas used for examining the association between community SESand primary outcomes following OHCA. We introduced potentialconfounding variables in a sequence of models. Risk factors weredirectly selected on the basis of previous literature and includedage, gender, initial ECG rhythm, witnessed status, bystander CPR,elapsed time interval such as call to ambulance arrival time to thescene, call to ambulance arrival time to ED in following each model.In Model 1, we adjusted for age group and gender. In Model 2, weadded initial ECG rhythm, witnessed status, and whether bystanderCPR was attempted. Finally in Model 3, we added elapsed timeintervals. The odds ratios (ORs) and 95% confidence intervals (95%CIs) were calculated.

3. Results

3.1. Demographic findings

During the study period, a total 36,724 of OHCA patients wereregistered in the CAVAS database. Of these, 2497 (6.8%) patientswere excluded due to lack of information on outcomes and/orgeographic information. The final sample available for analysisincluded 34,227 OHCAs. Table 1 shows the distributions of thecases and the denominator populations according to the level ofthe area deprivation index and urbanization category. The crudeincidence rate for OHCA per 100,000 population by district depri-vation level was 31.6 (Q1), 34.6 (Q2), 34.5 (Q3), 39.6 (Q4), and 46.7(Q5), respectively.

Table 2 shows the distribution of potential risk factors asso-ciated with outcomes of OHCA according to the area deprivationlevel where the OHCA occurred. There were no significant differ-ences in the distribution of age and gender of OHCA across differentlevels of area deprivation. However, the more deprived the dis-trict in which the OHCA occurred, the lower was the proportionsof arrests involving a shockable rhythm (p < 0.001), witnessed sta-tus (p < 0.001), and attempted bystander CPR (p < 0.001). Patientspresenting with OHCA in more deprived communities were lesslikely to receive CPR at the emergency department than patients inless deprived areas (p < 0.001). The elapsed time intervals showeda bi-modal distribution across levels of area deprivation. Q1 (theleast deprived) and Q4 or Q5 (most deprived) areas showed longerdelays (in minutes) for the response time and total transportationtime intervals.

Table 3 shows the distribution of our three outcomes of interest(return of spontaneous circulation; survival to hospital admission,

and survival to hospital discharge) according to the levels of areadeprivation. All three indicators of prognosis exhibited significantassociations with area deprivation (p < .0001). Patients experienc-ing an out-of-hospital cardiac arrest in the more deprived areaswere less likely to experience favorable prognosis (Fig. 1).

K.O. Ahn et al. / Resuscitation 82 (2011) 270–276 273

Table 2Demographic findings of risk factors according to deprivation index classification.

Variables Total Q1 (lowest) Q2 (low) Q3 (middle) Q4 (high) Q5 (highest) p-Value

N % N % N % N % N %

Total 34,227 6499 100.0 7868 100.0 7650 100.0 6490 100.0 5720 100.0Gender, total

Male 22,388 4188 64.4 5136 65.3 5017 65.6 4279 65.9 3768 65.90.06Female 11,839 2311 35.6 2732 34.7 2633 34.4 2211 34.1 1952 34.1

Age groupa

Pediatric 1071 228 3.5 251 3.2 223 2.9 225 3.5 144 2.50.81Adult 16,011 2953 45.8 3714 47.4 3711 4,8.71 3023 4,6.75 2610 45.7

Elderly 16,993 3272 50.7 3865 49.4 3684 4,8.36 3219 4,9.78 2953 51.7Initial rhythm

Non-shockable 27,329 5157 79.4 6355 80.8 6035 78.9 5162 79.5 4620 80.8<0.0001Shockable 1129 249 3.8 289 3.7 272 3.6 172 2.7 147 2.6

Unknown 5769 1093 16.8 1224 15.6 1343 17.6 1156 17.8 953 16.7Witnessed status

No 20,498 3792 58.4 4567 58.1 4590 60.0 3986 61.4 3563 62.3<0.0001Yes 13,729 2707 41.7 3301 42.0 3060 40.0 2504 38.6 2157 37.7

Bystander CPRNo 12,562 1835 28.2 2698 34.3 2638 34.5 2789 43.0 2602 45.5

<0.0001Yes 478 119 1.8 114 1.5 120 1.6 67 1.0 58 1.0Unknown 21,187 4545 69.9 5056 64.3 4892 64.0 3634 56.0 3060 53.5

Cause of arrestCardiogenic 18,971 3512 54.0 4461 56.7 4344 56.8 3485 53.7 3169 55.4

0.75Non-cardiogenic 15,256 2987 46.0 3407 43.3 3306 43.2 3005 46.3 2551 44.6CPR at EMS

No 9118 1685 25.9 2089 26.6 2195 28.7 1652 25.5 1497 26.20.78Yes 25,109 4814 74.1 5779 73.5 5455 71.3 4838 74.6 4223 73.8

CPR at EDNo 20,363 3646 56.1 4630 58.9 4506 58.9 3936 60.7 3645 63.7

<0.0001Yes 13,864 2853 43.9 3238 41.2 3144 41.1 2554 39.4 2075 36.3Elapsed intervals, min

Call to scene arrival 34,227 8.1* 7.4 7.6** 6.9 7.7** 5.7 7.9* 6.4 7.8* 8.6 <0.0001Call to ED arrival 34,227 25.4* 16.4 23.5** 15.0 23.7** 15.8 24.3*** 15.1 24.8*** 17.3 <0.0001

CPR: cardiopulmonary resuscitation; EMS: emergency medical service; ED: emergency department.Notice: The Cochran–Armitage test was used for statistical analysis for risk factors except elapsed intervals which were tested by the ANOVA test and post hoc comparison.N ber s

natora

3

sdmcdtt

aa

TO

o significant difference is between same number of asterisk. Different asterisk numa The number of patient with unknown age was 152, which was used as denomi

dult (15 ≤ age < 65 years), and elderly (age ≥ 65 years), respectively.

.2. Multivariate logistic regression analysis

Table 4 shows the results of the multivariate logistic regres-ion analysis, where we regressed survival to discharge on the areaeprivation level (reference group = least deprived areas, Q1). Inodel 1 (adjusting for age group and gender), there was a signifi-

ant association between area deprivation and survival to hospitalischarge. Compared to patients in the least deprived districts (Q1),hose in the most deprived districts (Q5) experienced roughly half

he survival (OR 0.57, 95% CI: 0.45–0.72).

When we added initial ECG rhythm, attempted bystander CPR,nd witnessed status in model 2, the differences were slightlyttenuated, but remained statistically significant. Finally, when we

able 3utcomes according to deprivation index classification.

Outcomes Total Q1 (lowest) Q2 (low)

N % N % N %

Total 34,227 100.0 6499 100.0 7868 100.0Any return of spontaneous circulation

No 28,760 84.0 5298 81.5 6541 83.1Yes 5467 16.0 1201 18.5 1327 16.9

Survival to hospital admissionNo 30,248 88.4 5619 86.5 6844 87.0Yes 3979 11.6 880 13.5 1024 13.0

Survival to hospital dischargeNo 33,398 97.6 6297 96.9 7650 97.2Yes 829 2.4 202 3.1 218 2.8

a Cochrane–Armitage test for trend was used.

hows the significant difference comparing with other groups.for calculation of proportion. Age group was classified as pediatric (age < 15 years),

adjusted for elapsed time intervals before reaching treatment (inmodel 3), the pattern still persisted, with patients in the mostdeprived districts experiencing roughly half the odds of survivalcompared to patients in the least deprived districts. Our interpre-tation of these models is that the disparities in survival betweenless vs. more deprived areas could not be explained by out-of-hospital factors, such as attempted CPR by bystanders, or delaybefore reaching the emergency room. These did not mediate theassociation between area deprivation and our primary outcome

variable (survival to hospital discharge). In turn, this suggests thatsurvival to hospital discharge is likely to be explained by differ-ences in the quality of treatment within hospitals serving differentareas.

Q3 (middle) Q4 (high) Q5 (highest) p-Valuea

N % N % N %

7650 100.0 6490 100.0 5720 100.0

6431 84.1 5511 84.9 4979 87.1<0.00011219 15.9 979 15.1 741 13.0

6782 88.7 5819 89.7 5184 90.6<0.0001868 11.4 671 10.3 536 9.4

7469 97.6 6364 98.1 5618 98.2<0.0001181 2.4 126 1.9 102 1.8

274 K.O. Ahn et al. / Resuscitation 82 (2011) 270–276

Table 4Odds ratios (ORs) with 95% confidence intervals (CIs) from multivariate logistic regression models for survival to hospital discharge.

Variables Model 1 Model 2 Model 3

Adjusted OR 95% CI Adjusted OR 95% CI Adjusted OR 95% CI

Deprivation indexQ1 (lowest) 1.00 1.00 1.00Q2 (low) 0.88 0.72 1.07 0.91 0.75 1.11 0.84 0.69 1.03Q3 (middle) 0.74 0.61 0.91 0.77 0.62 0.94 0.72 0.59 0.89Q4 (high) 0.61 0.49 0.77 0.65 0.52 0.82 0.61 0.49 0.77Q5 (highest) 0.57 0.45 0.72 0.62 0.48 0.79 0.58 0.45 0.74

GenderMale 1.00 1.00 1.00Female 0.95 0.82 1.11 0.98 0.84 1.14 0.94 0.8 1.09

Age groupa

Pediatric 0.92 0.64 1.32 0.92 0.64 1.33 0.84 0.58 1.22Adult 1.00 1.00 1.00Elderly 0.50 0.43 0.58 0.50 0.43 0.58 0.52 0.44 0.6

Initial rhythmNon-shockable 1.00 1.00Shockable 1.1 1.08 1.12 1.08 1.06 1.1

Witnessed statusNo 1.00 1.00Yes 3.13 2.7 3.63 3 2.59 3.49

Bystander CPRNo 1.00 1.00Yes 1.04 1.02 1.06 1.03 1.02 1.05

Elapsed intervals, (unit = 1.0 min)Call to scene arrival 0.97 0.94 0.99Call to ED arrival 0.96 0.95 0.97

CPR: cardiopulmonary resuscitation; ED: emergency department.Notice: Gender and age group was adjusted in Model 1. Initial rhythm, witnessed status, and bystander CPR was added in Model 2 and finally elapsed time intervals addedin Model 3.

a The number of patient with unknown age was 152, which was used as denominatoradult (15 ≤ age < 65 years), and elderly (age ≥ 65 years), respectively.

Fto

4

iAisiesid(t

ig. 1. Outcomes according to area deprivation levels. *Cochrane–Armitage test forrend showed statistically significant association between outcomes and increasef area deprivation level.

. Discussion

To our knowledge, this is the first study of area variationsn outcomes following out-of-hospital cardiac arrests in Korea.lthough socioeconomic status has been studied extensively as an

ndividual-level predictor of cardiovascular disease outcomes, ourtudy adds two new dimensions to this association. First, we exam-ned out-of-hospital cardiac arrests which have not been examinedxtensively in the literature. Secondly, we examined area-level

ocioeconomic status as a determining factor in prognosis follow-ng OHCA. Even though South Korea does not suffer from the sameegree of socio-economic inequality compared to the United Statesfor example, every Korean has access to health care, and the dis-ribution of income and education is more egalitarian than in the

for calculation of proportion. Age group was classified as pediatric (age < 15 years),

US), our study nonetheless documented a twofold disparity in theoutcome of OHCA comparing individuals who had their event in theleast deprived districts compared to those who experienced theirarrest in the most deprived district.

In the US, Cowieet al. reported that the race was a major pre-dictor of survival following OHCA.14 Initial ECG findings differed byrace in their study, and resulted in outcome disparities. In anotherU.S. study, Brookoff et al. reported that African Americans were lesslikely to receive CPR from bystanders compared to whites.15 Beckeret al. reported similar findings.9 Even though they adjusted forpotential predictors of survival to discharge, black race remaineda significant risk factor for poor prognosis. These observed dis-parities by race/ethnic status in the United States may be due tothe fact that black Americans are more likely to live in segregatedand deprived neighborhoods compared to white Americans. Conse-quently, when they suffer a cardiac arrest on the street, bystandersmay be less “health literate”, i.e. be equipped with the skills to per-form CPR. In the study by Iwashyna et al. patients with arrests inracially integrated neighborhoods were most likely to be providedwith CPR, followed by those in predominately white neighbor-hoods, with the lowest rates of CPR provision in predominatelyblack neighborhoods.10 By contrast, neither the socioeconomic sta-tus, number of elderly, nor the occupational characteristics of theneighborhood appeared to influence CPR provision.

Clarke et al. reported a significant association between lowertax-assessed property value and lower survival to dischargewhile no significant association between lower median householdincome and poor outcome was there.16 Galea et al. examined car-diac arrests in New York and found no association between median

household income of area and survival after OHCA, or EMS responsetime interval. They concluded that the disparity was minimaland did not affect survival following OHCA.8 The lower house-hold level SES was not consistent for the associated with worseoutcome.

citatio

batai(lrittuhts

gdtctclfaeptSt

5

airalm

aTdOTsaalpdhas

6

hpsfim

1

1

1

1

1

1

1

1

1

1

K.O. Ahn et al. / Resus

There are competing hypotheses for how area-level SES maye linked with prognosis after OHCA. On the one hand, having anrrest in an area with lower levels of health literacy may reducehe chances of attempted bystander CPR. Our study confirmed thisssociation, with almost twice as higher rate of attempted CPRn the least deprived areas compared to the most deprived areasTable 2). On the other hand, if coronary disease is less common iness deprived areas, residents and EMS services may be slower toespond compared to areas where OHCA is more common. Our find-ng of a bimodal distribution of response times is consistent withhis hypothesis. An alternative explanation for the longer delay inhe least deprived areas is that these are the most densely pop-lated metropolitan areas in the country where residents live inigh-rise apartment buildings with congested traffic – both fac-ors which contribute to significant delays in reaching emergencyervices.17–20

Regardless of actions taken outside the hospital, our study sug-ests that OHCA prognosis – measured by survival to hospitalischarge – is twice as good in the least deprived areas comparedo the most deprived areas. This pattern suggests that quality ofare inside the hospital is a decisive factor. For example, therapeu-ic hypothermia, optimal critical care, and emergency percutaneousoronary intervention critically determine the chain of survival fol-owing OHCA.21,22 These three major components for outcomesrom OHCA may depend on the level of community resources. Welso found that there was a disparity in resuscitation efforts at themergency department according to area deprivation level. Thisattern echoes previous findings in the literature which suggesthat socially disadvantaged groups (minority racial groups, lowES groups) receive less intensive treatments such as reperfusionherapy for acute myocardial infarction.23–25

. Limitations

We enrolled all causes of OHCA in our analysis. However, cardiacrrest occurring in the context of trauma has a different pathophys-ology compared with underlying coronary disease. This may haveesulted in a misclassification of the association between area SESnd OHCA outcomes. We excluded 2497 (6.8%) of patients due toack of information on outcomes and address. These excluded data

ight make bias although the proportion was a little.The information on bystander CPR was very incomplete with

very high proportion assigned to the unknown group (61.9%).his incomplete information might have biased our findings. Weid not analyze the association between individual level SES andHCA outcomes, as we lacked information on individual SES.his study was done in a basic to intermediate service level EMSetting where a few treatment options such as laryngeal maskirway, intravenous fluid infusion (intermediate level) as well asutomatic external defibrillation or cardiac compression (basicevel) by EMS providers are allowed. In this context, there is noublic access defibrillator program, designated regionalized car-iac arrest center, or specialized centers capable of therapeuticypothermia for post-resuscitation care. These study setting char-cteristics limit the generalizability of our study findings to otherettings.

. Conclusion

Area deprivation was significantly associated with survival to

ospital discharge among OHCA patients when we adjusted forotential confounders (age, gender, initial ECG rhythm, witnessedtatus, bystander CPR, and elapsed time interval) in Korea. Ourndings call for additional studies to examine the mechanismsediating this association, and point to the possibility of improving

2

2

n 82 (2011) 270–276 275

the distribution of emergency-service resources across the countryto address this health disparity.

Conflict of interest disclosures

This study was financially supported by the Korea Center forDisease Control and Prevention in 2009. All authors are not relatedwith any other conflicts of interest in this study.

Acknowledgements

The Korea Center for Disease Control and Prevention financiallysupported this study. Authors appreciate Dr. Mia Son for providingher knowledge regarding Carstair index calculation.

References

1. Rosamon W, Flegal K, Furie K, et al. Heart disease and stroke statistics: 2008updates: a report from the American Heart Association Statistics Committeeand Stroke Statistics Subcommittee. Circulation 2008;117:e25–146.

2. Jacobs I, Nadkarni V, Bahr J, et al. International Liaison Committee onResuscitation; American Heart Association; European Resuscitation Council;Australian Resuscitation Council; New Zealand Resuscitation Council; Heartand Stroke Foundation of Canada; InterAmerican Heart Foundation; Resusci-tation Councils of Southern Africa; ILCOR Task Force on Cardiac Arrest andCardiopulmonary Resuscitation Outcomes.Cardiac arrest and cardiopulmonaryresuscitation outcome reports: update and simplification of the Utstein tem-plates for resuscitation registries: a statement for healthcare professionals froma task force of the International Liaison Committee on Resuscitation (Ameri-can Heart Association, European Resuscitation Council, Australian ResuscitationCouncil, New Zealand Resuscitation Council, Heart and Stroke Foundation ofCanada, InterAmerican Heart Foundation. Resuscitation Councils of SouthernAfrica). Circulation 2004;110:3385–97.

3. Nichol G, Thomas E, Callaway CW, et al. Stiell I; Resuscitation Outcomes Consor-tium Investigators. Regional variation in out-of-hospital cardiac arrest incidenceand outcome. JAMA 2008;300:1423–31.

4. Mensah GA, Mokdad AH, Ford ES, Greenlund KJ, Croft JB. State of disparities incardiovascular health in the United States. Circulation 2005;111:1233–41.

5. Song YM, Ferrer RL, Cho SI, Sung J, Ebrahim S, Davey Smith G. Socioeconomic sta-tus and cardiovascular disease among men: the Korean national health serviceprospective cohort study. Am J Public Health 2006;96:152–9.

6. Hallstrom A, Boutin P, Cobb L, Johnson E. Socioeconomic status and predictionof ventricular fibrillation survival. Am J Public Health 1993;83:245–8.

7. Sayegh AJ, Swor R, Chu KH, et al. Does race or socioeconomic status predictadverse outcome after out of hospital cardiac arrest: a multi-center study. Resus-citation 1999;40:141–6.

8. Galea S, Blaney S, Nandi A, et al. Explaining racial disparities in incidence of andsurvival from out-of-hospital cardiac arrest. Am J Epidemiol 2007;166:534–43.

9. Becker LB, Han BH, Myer PM, et al. Racial differences in the incidence of car-diac arrest and subsequent survival. The CPR Chicago Project. N Engl J Med1993;329:600–6.

0. Iwashyna TJ, Christakis NA, Becker LB. Neighborhoods matter: a population-based study of provision of cardiopulmonary resuscitation. Ann Emerg Med1999;34:459–68.

1. Park CB, Shin SD, Suh GJ, et al. Pediatric out-of-hospital cardiac arrest in Korea:a nationwide population-based study. Resuscitation 2010;81:512–7.

2. Cha WC, Shin SD, Song KJ, Jung SK, Suh GJ. Effect of an independent-capacityprotocol on overcrowding in an urban emergency department. Acad Emerg Med2009;16:1277–83.

3. Carstairs V, Morris R. Deprivation: explaining differences in mortality betweenScotland and England and Wales. BMJ 1989;299:886–9.

4. Cowie MR, Fahrenbruch CE, Cobb LA, Hallstrom AP. Out-of-hospital car-diac arrest: racial differences in outcome in Seattle. Am J Public Health1993;83:955–9.

5. Brookoff D, Kellermann AL, Hackman BB, Somes G, Dobyns P. Do blacks getbystander cardiopulmonary resuscitation as often as whites? Ann Emerg Med1994;24:1147–50.

6. Clarke SO, Schellenbaum GD, Rea TD. Socioeconomic status and survival fromout-of-hospital cardiac arrest. Acad Emerg Med 2005;12:941–7.

7. Campbell JP, Gratton MC, Salomone 3rd JA, Watson WA. Ambulance arrival topatient contact: the hidden component of prehospital response time intervals.Ann Emerg Med 1993;22:1254–7.

8. Morrison LJ, Angelini MP, Vermeulen MJ, Schwartz B. Measuring the EMS patientaccess time interval and the impact of responding to high-rise buildings. PrehospEmerg Care 2005;9:14–8.

9. Silverman RA, Galea S, Blaney S, et al. The “vertical response time”: barriers to

ambulance response in an urban area. Acad Emerg Med 2007;14:772–8.

0. Becker LB, Ostrander MP, Barrett J, Kondos GT. Outcome of CPR in a largemetropolitan area where are the survivors? Ann Emerg Med 1991;20:355–61.

1. Spaite DW, Stiell IG, Bobrow BJ, et al. Effect of transport interval on out-of-hospital cardiac arrest survival in the opals study: implications for triagingpatients to specialized cardiac arrest centers. Ann Emerg Med 2009;54:248–55.

2 citatio

2

2

76 K.O. Ahn et al. / Resus

2. Sunde K, Pytte M, Jacobsen D, et al. Implementation of a standardised treat-ment protocol for post resuscitation care after out-of-hospital cardiac arrest.Resuscitation 2007;73:29–39.

3. Bradley EH, Herrin J, Wang Y, et al. Racial and ethnic differences in time to acutereperfusion therapy for patients hospitalized with myocardial infarction. JAMA2004;292:1563–72.

2

2

n 82 (2011) 270–276

4. Peterson ED, Shah BR, Parsons L, et al. Trends in quality of care for patients withacute myocardial infarction in the National Registry of Myocardial Infarctionfrom 1990 to 2006. Am Heart J 2008;156:1045–55.

5. Jneid H, Fonarow GC, Cannon CP, et al. Get with the guidelines steering com-mittee and investigators. Sex differences in medical care and early death afteracute myocardial infarction. Circulation 2008;118:2803–10.