Indexed in MEDLINE, SCI & KoreaMed
pISSN 1011-8934   eISSN 1598-6357
Open Access, Peer-reviewed, Weekly
    J Korean Med Sci. 2024 Mar 11;39(9):e86. English.
    Published online Feb 21, 2024.
    https://doi.org/10.3346/jkms.2024.39.e86
    © 2024 The Korean Academy of Medical Sciences.
    Original Article

    Spatiotemporal Analysis of Out-of-Hospital Cardiac Arrest Incidence and Survival Outcomes in Korea (2009–2021)

    Naae Lee,1 Seungpil Jung,1 Young Sun Ro,2,3 Jeong Ho Park,2,3 and Seung-sik Hwang1,2
      • 1Department of Public Health Sciences, Graduate School of Public Health, Seoul National University, Seoul, Korea.
      • 2Laboratory of Emergency Medical Services, Seoul National University Hospital Biomedical Research Institute, Seoul, Korea.
      • 3Department of Emergency Medicine, Seoul National University Hospital, Seoul, Korea.
    • Address for Correspondence: Seung-sik Hwang, MD, PhD. Department of Public Health Sciences, Seoul National University Graduate School of Public Health, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, Korea. Email: cyberdoc@snu.ac.kr
    Received August 20, 2023; Accepted January 09, 2024.

    This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

    Abstract

    Background

    Out-of-hospital cardiac arrest is a major public health concern in Korea. Identifying spatiotemporal patterns of out-of-hospital cardiac arrest incidence and survival outcomes is crucial for effective resource allocation and targeted interventions. Thus, this study aimed to investigate the spatiotemporal epidemiology of out-of-hospital cardiac arrest in Korea, with a focus on identifying high-risk areas and populations and examining factors associated with prehospital outcomes.

    Methods

    We conducted this population-based observational study using data from the Korean out-of-hospital cardiac arrest registry from January 2009 to December 2021. Using a Bayesian spatiotemporal model based on the Integrated Nested Laplace Approximation, we calculated the standardized incidence ratio and assessed the relative risk to compare the spatial and temporal distributions over time. The primary outcome was out-of-hospital cardiac arrest incidence, and the secondary outcomes included prehospital return of spontaneous circulation, survival to hospital admission and discharge, and good neurological outcomes.

    Results

    Although the number of cases increased over time, the spatiotemporal analysis exhibited a discernible temporal pattern in the standardized incidence ratio of out-of-hospital cardiac arrest with a gradual decline over time (1.07; 95% credible interval [CrI], 1.04–1.09 in 2009 vs. 1.00; 95% CrI, 0.98–1.03 in 2021). The district-specific risk ratios of survival outcomes were more favorable in the metropolitan and major metropolitan areas. In particular, the neurological outcomes were significantly improved from relative risk 0.35 (0.31–0.39) in 2009 to 1.75 (1.65–1.86) in 2021.

    Conclusion

    This study emphasized the significance of small-area analyses in identifying high-risk regions and populations using spatiotemporal analyses. These findings have implications for public health planning efforts to alleviate the burden of out-of-hospital cardiac arrest in Korea.

    Graphical Abstract

    Keywords
    Incidence; Out-of-Hospital Cardiac Arrest; Spatiotemporal Analysis; Bayesian; Survival

    INTRODUCTION

    Out-of-hospital cardiac arrest (OHCA) is a major cause of morbidity and mortality worldwide.1, 2 Its incidence and survival rates vary considerably across geographic regions, underscoring its significance as a global health issue.3 OHCA incidence is reportedly 55 per 100,000 people in North America,4 with an 8–12% survival rate at hospital discharge5; however, the OHCA incidence differs across studies and regions. In the Republic of Korea, OHCA is a significant public health concern owing to the aging population,6, 7 with an estimated incidence of 64.7 cases per 100,000 person-years in 2021.8 Despite advancements in emergency medical services (EMSs) and resuscitation protocols, the overall survival rates for patients with OHCA remain low (global average < 10%).9 Furthermore, there is a marked difference in the incidence and survival outcomes of OHCA globally.10 Consequently, understanding the spatial and temporal aspects of the incidence and survival outcomes of OHCA is crucial for developing targeted interventions and effective strategies to prevent and treat cardiac arrest in Korea.

    Improving survival rates requires strengthening each link in the chain of survival, particularly regarding the initial interventions by bystanders, i.e., witnessing the event, immediate emergency activation, early cardiopulmonary resuscitation (CPR), and utilization of automated external defibrillators (AEDs).11, 12 Factors influencing OHCA survival can be classified into spatial and non-spatial; however, much is unknown regarding their relative importance in improving survival. Furthermore, both factors need to be considered simultaneously.13 A spatiotemporal analysis is a valuable tool for public health interventions that can be used to identify sources of heterogeneity in health risks and outcomes.14 A model has been developed to explain and predict the spatiotemporal interactions and demographic characteristics of OHCA.15, 16 Moreover, studies employing spatiotemporal analysis provide policymakers and EMS providers with evidence to support policy decision-making.17 A previous comparison of OHCA incidence and survival outcomes between patients in Singapore and Victoria, Australia, revealed survival disparities that could be attributed to variations in EMSs.18

    Besides investigations on the overall occurrence and spatial distribution of OHCA, studies incorporating both temporal and spatial dimensions have explored aspects such as the spatial cluster of bystander CPR (BCPR),19 optimization of AED deployment,20 drone-assisted AED delivery,21 and EMS response time within the chain of survival.13, 17, 22 Conducting spatiotemporal evaluations across a broad geographical context offers advantages such as identifying high-risk areas at the small-area level, addressing regional disparities,14 and optimizing the allocation of limited resources, which maximizes the provision of community-based educational programs.23 However, studies have mainly focused on the OHCA incidence,14, 15, 16, 23 identifying risk factors associated with improved survival rates,17, 22 spatial variations,13 or the role of bystanders and distribution of AEDs.20

    Despite a growing interest in the geospatial analysis of OHCA, there is limited research on its spatiotemporal epidemiology in Korea. Previous studies have examined the temporal patterns of the domestic incidence rates of OHCA by sex24 and incidence rates with respect to the night-time residential population and daytime transient population.25 However, there is a lack of studies applying spatiotemporal analysis to investigate the occurrence and outcomes of OHCA comprehensively. Therefore, this study aimed to bridge this gap by conducting a spatiotemporal analysis of the OHCA incidence and survival outcomes in Korea, utilizing Bayesian hierarchical models and the Korean OHCA registry.

    METHODS

    Study design and setting

    This cross-sectional study used data from the Korean OHCA registry provided by the Korea Disease Control and Prevention Agency (KDCA). The Republic of Korea occupies an area of 100,210 km2, with a population exceeding 51 million. It comprises 250 administrative districts, including six metropolitan cities (Gwangyeoksi), one special self-governing city (Teukbyeol jachisi), one special city (Teukbyeolsi), and nine provinces (do) including one special self-governing province (Teukbyeol jachido). Metropolitan cities encompass both regional and local governance roles. At the regional scale, there are 250 districts designated as cities (si), counties (gun), or districts (gu). These districts are then subdivided into 3,500 administrative divisions such as eup (town), myeon (township), dong? (neighborhood).26 Detailed local information is presented in Supplementary Fig. 1. Based on the reference year of 2021 (250 administrative districts), the integrated districts were aggregated by summing their respective counts, while the divided districts were calculated by dividing them by 1/N. Each county is served by dedicated healthcare and administrative authorities operating within their respective jurisdictions.

    The Korean EMS operates as a government-based system,25 delivering prompt and efficient medical care during emergencies. The Korean EMS, which has regional headquarters in the National Fire Agency, employs a partially dual dispatch response system; it uses basic life support fire engines and advanced cardiovascular life support ambulances for suspected cases of OHCA.27 Prehospital EMS providers within this system are trained emergency medical technicians and nurses affiliated with the fire department. All patients receive CPR following the EMS protocol and are transported to the nearest emergency department.28

    Data collection and study population

    The nationwide Korean OHCA registry was constructed by the KDCA in collaboration with the Central Fire Services in 2006. It is being updated in cooperation with the National Fire Administration, 17 city and provincial fire departments, and 712 hospitals nationwide.29 Trained experts in medical record reviewing, specializing in the application of Utstein guidelines, were employed to systematically review medical records, focusing on variables associated with the etiology, risks, and outcomes of OHCA.30, 31 The survey questions were adapted and refined based on the international OHCA investigation criteria outlined in the Utstein Style and the Resuscitation Outcome Consortium (ROC) Project, taking into consideration the practicality of data collection in the domestic context.8 The information extracted from the Korean OHCA registry comprised various aspects, including the date of OHCA, demographic details of the individuals affected, location of the OHCA incident, presence of witnesses, bystander CPR, causes of OHCA, initial electrocardiographic rhythm, survival to discharge, and CPC upon discharge.29

    This study included all EMS-assessed OHCA cases recorded in Korea between January 1, 2009, and December 31, 2021. The annual population data for 250 administrative districts, including age and sex distributions for each year, were obtained from Statistics Korea.

    Outcome variables and measurements

    We collected the following patient information: 1) demographic information—sex (male or female) and age (continuous); 2) prehospital factors—past medical history (hypertension, diabetes mellitus, heart disease, chronic kidney disease, respiratory, stroke, and dyslipidemia), cardiac etiology (medical or non-medical), witness status, place of OHCA (public or private), and type of bystander interventions (whether the patient received CPR or AED); and 3) survival outcomes—prehospital return of spontaneous circulation (ROSC), survival to hospital admission, survival to hospital discharge, and good neurological recovery (cerebral performance category [CPC] score of 1 or 2 at discharge). A CPC score of 1 indicated mild or no neurological deficits, whereas a score of 2 indicated moderate cerebral disability.24

    For bystander interventions such as CPR and AED, data were not available for the whole period; hence, only data from 2013 to 2021 were analyzed. To compare regional characteristics, we used population density and the aging index from the Korean Statistical Information Service. Although the data were surveyed every 5 years from 2000 and annually from 2016, only data from 2010, 2015, and 2020 were used in this study. Population density is the number of people divided by land area in square kilometers. The aging index refers to the number of older individuals (aged 65 years and over) per 100 individuals younger than 14 years in a specific population.

    We investigated possible risk factors related to the incidence and survival outcomes of OHCA. For incidence, we used indicators available in community health surveys for hypertension and diabetes diagnosis rates (age 30 years and above); for survival outcomes, indicators for CPR education experience and AED mannequin practice experience were used. As additional information in the community health survey data, the CPR training experience rate was investigated every 2 years from 2012 (with the latest investigation in 2020) and AED mannequin practice experience rate was investigated only once, in 2020.

    To observe the overall temporal change, the 13 years were divided into three intervals (Phase 1: 2009–2012, Phase 2: 2013–2016, and Phase 3: 2017–2021) for demographic information. For most results, the first year was 2009, the middle year was 2015, and the last year was 2021.

    Statistical analysis

    The standardized incidence ratio (SIR) is a commonly used epidemiological measure to compare the incidence of a particular disease in a study population with that in a reference population.32 We evaluated the distribution of OHCA incidence using the SIR at the district-level stratified for sex and age in groups (0–9, 10–19, 20–29, …, 100+). The number of observed OHCA cases was calculated based on the locations of incident occurrences at the district level. The SIR was calculated as the ratio of the observed number of OHCA cases to the expected number of cases in each year and each district area and presented as the 95% credible interval (95% CrI).

    We employed an intercept-only Bayesian spatiotemporal model based on the Integrated Nested Laplace Approximation (INLA)33, 34 to assess the spatial and temporal risks of survival outcomes.35 The model included the Besag–York–Mollie model structure and a nonparametric time trend and spatiotemporal interaction. To find the best-fit model, we used a model selection approach using the Watanabe–Akaike information criterion (WAIC) among four likelihood distributions, i.e., Poisson, negative binomial, zero-inflated Poisson, and zero-inflated negative binomial. The model with the lowest WAIC was used for each outcome variable, and a Poisson spatiotemporal model fit the best with our data (Supplementary Table 1). To express spatial heterogeneity, we calculated the relative risk (RR) and corresponding 95% CrI. RR represents whether area i exhibits a higher RR (RRi>1) with a high risk, respectively. Areas with RRi=1 indicated the same risk as that expected from the standard population.36

    Additionally, we computed the posterior probability of the spatial effects greater than a given threshold value to represent the uncertainty associated with RRi.37 An exceedance probability (EP) of ≥ 80% based on the Richardson criterion indicated a high certainty of surpassing the RRi thresholds in the district.38 The results are presented as EPs to investigate individual areas with a high risk of evolving. The posterior EP serves as a useful tool for detecting hotspots, which refer to isolated areas exhibiting a typically high RR.39

    Detailed formulas and explanations are provided in the supplementary material with the result of the comparison model (Supplementary Data 1 and Supplementary Table 1). All statistical analyses were performed in R (version 4.2.3; The R Foundation for Statistical Computing, Vienna, Austria),40 and the Bayesian spatiotemporal modeling was performed using the R-INLA package (version 22.12.16).34

    Ethics statement

    This study was reviewed and approved by the Institutional Review Board (IRB) of Seoul National University (IRB No. E2303/004-004). The requirement for informed consent was waived by the IRB.

    RESULTS

    Demographics of OHCA cases and survival outcomes

    We analyzed 368,480 OHCA cases and assessed their survival outcomes between January 1, 2009, and December 31, 2021. The basic characteristics of the patients are summarized and divided into three intervals (Phase 1: 2009–2012, Phase 2: 2013–2016, and Phase 3: 2017–2021) as shown in Table 1. Overall, the total number of OHCA cases increased compared with that in previous phases (22,693 patients in 2009 to 33,041 patients in 2021). Approximately 64.2% of cases involved male patients, and the mean age increased from 62.2 to 66.7 years (64.9 years over the entire study period).

    Table 1
    Characteristics of out-of-hospital cardiac arrest in Korea from 2009 to 2021 considered in three phases

    Regarding the etiology of cardiac arrest, medical causes comprised the underlying etiology in 74.0% of the patients; among them, cases of 69.1% of patients were attributable to cardiogenic causes. In non-medical cases (injury), proportions of transport accidents decreased from 9.6% to 6.6%; however, the incidence of fall accidents increased from 3.8% to 5.0% in Phase 3. Regarding the medical history of patients with OHCA, hypertension was the most prevalent (26.7%), followed by diabetes (17.5%) and heart disease (2.1%).

    Analysis of cardiac arrest incidents revealed a significant increase in the proportion of cases witnessed (37,866 cases in Phase 1 to 76,287 cases in Phase 3). The incidence of cardiac arrest was higher in non-public locations, specifically private homes and nursing homes, and accounted for 64% of all cases. Among them, 77% of non-public cases occurred in residential houses. The recognition of cardiac arrest incidents improved significantly, resulting in a substantial increase in the use of BCPR by a layperson from 43.0% (49,999 cases) in Phase 1 to 51.7% (79,407 cases) in Phase 3. However, the use of AEDs, which accounted for 2.4% of all cases, was lower than that of BCPR.

    Analysis of survival outcomes revealed improvements in prehospital ROSC, survival to hospital admission and discharge, and good neurological recovery. Detailed tables presenting survival outcomes are provided, categorizing men and women in age groups (Supplementary Tables 2 and 3). Regarding sex differences, survival rates were consistently 2–4 times higher in men than in women across all survival outcomes. Regarding age, men and women showed the highest incidence of cardiac arrest in their 70s. Notably, relatively high survival and recovery rates were observed among patients in their 50s and 60s.

    SIR

    Fig. 1A illustrates the district-level maps depicting the SIR for the OHCA incidence across three years (2009, 2015, and 2021). To compare regional characteristics related to the occurrence of OHCA, the population density is presented in Fig. 1B and the aging index is presented in Fig. 1C. We identified that there was no significant difference in population density from 2010 to 2021, and it was confirmed that the population was concentrated in the metropolitan areas and major metropolitan areas. Conversely, in the aging index, populations in the metropolitan areas have younger ages, while those in the rural areas are gradually aging. Comparing the incidence of OHCA and population characteristics, analysis of the data from 2009 onward revealed a discernible temporal trend for the SIR, with a gradual decrease over time. The SIR of OHCA in 250 regions was 1.00 (95% CrI, 0.98–1.03) in 2021, 1.07 (95% CrI, 1.04–1.09) in 2009, and 1.12 (95% CrI, 1.09–1.14) in 2015. In 2009, 64% of districts exhibited an SIR < 1, whereas the corresponding value was only 52.4% in 2021, which can imply that the incidence of OHCA is decreasing at the district level. Over time, a decrease in the SIR was observed in provinces where major cities were located. Specifically, in regions characterized by high population density, such as metropolitan areas and major urban centers, there was no notable elevation in the occurrence of OHCA. Notably, the Northeastern (Gangwon) and Southern (Jeju) regions displayed higher SIR values than the other regions. A similar trend for the aging index map (Fig. 1B) was observed in these regions, particularly in areas outside the metropolitan areas with higher aging indices. In a similar pattern, when examining the map of the diagnosis experience rate of hypertension and diabetes over the age of 30 years, the prevalence rate increased over time, and it was confirmed that the tendency of cardiac arrest increased in the Eastern region with a high prevalence rate (Supplementary Fig. 2).

    Fig. 1
    Standardized incidence ratios, population densities and aging indices at the district level.

    Moreover, as a substantial number of cases (74%) with a cardiac etiology fall under the medical category, we illustrated the mapping of SIR exclusively for medically classified cases (Supplementary Fig. 3). In 2009, the central (Chungcheong Province), northeastern (Gangwon Province), and certain southwestern areas (Jeolla Province) showed slightly higher SIR compared to the overall OHCA cases. However, comparing 2015 and 2021, the areas with high overall cardiac arrest rates also exhibited similarly high rates for medical cases. Examining the spatial trends from 2009 to 2021, a visual inspection revealed a decreasing trend in the SIR, consistent with the overall analysis result of OHCA cases.

    Temporal, district-specific RR and EP

    Fig. 2 displays time trend curves illustrating the standardized RR stratified by sex and age groups from 2009 to 2021. With regards to the RR for OHCA, the value fluctuated around 1.0; from 2013 to 2015, it showed a slight increase, but from 2016 onward, the observed RR was < 1.0. Since 2015, significant improvements were observed in key outcomes among OHCA survivors, such as neurological recovery and prehospital ROSC.

    Fig. 2
    Temporal plot of relative risk of OHCA incidence and survival outcomes from 2009 to 2021.
    OHCA = out-of-hospital cardiac arrest, ROSC = return of spontaneous circulation.

    Regarding the district-specific RR, Fig. 3 presents the survival outcomes of OHCA cases for 3 years, and maps of all years are presented in Supplementary Figs. 4, 5, 6, 7. In addition, the distribution of the CPR education experience and AED mannequin practice experience, for the comparison of regional characteristics according to survival results, is presented in Supplementary Fig. 8.

    Fig. 3
    Geographical distribution map of the relative risk of out-of-hospital cardiac arrest. survival outcomes in the Republic of Korea, 2009–2021. (A) Prehospital return of spontaneous circulation. (B) Survival to hospital admission. (C) Survival to hospital discharge. (D) Good neurological recovery.

    RR of prehospital ROSC over all districts ranged from 0.38 (95% CrI, 0.35–0.42) in 2009 to 1.65 (95% CrI, 1.58–1.73) in 2021 (Fig. 3). Throughout the study period, the prehospital ROSC was high in the northwestern regions (e.g., Incheon, part of Gyeonngi-do). Meanwhile, survival to hospital admission and discharge rates were high in metropolitan areas and major metropolitan areas (e.g., Gwangju and Daejeon). The most improved survival outcome was neurological recovery (RR increased from 0.35 [95% CrI, 0.31–0.39] in 2009 to 1.75 [95% CrI, 1.75–1.86] in 2021). We found an RR above 1 in 5 of the 250 districts in 2009, but in 202 of the 250 districts in 2021. Moreover, the highest RR value was observed in the capital city (Seoul), ranging from 2.79 (95% CrI, 1.95–3.88) to 5.65 (95% CrI, 4.11–7.60).

    The CPR training experience rate was on the rise geographically, and the AED practice experience rate is still at a very low level regionally. When these factors and survival results were compared, survival results improved somewhat in areas with a high rate of CPR education experience but were not clearly revealed.

    Fig. 4 illustrates the posterior probability of exceeding the RR threshold for the 250 districts, with regional emergency medical centers in the area marked as black dots. The EPs of an RR > 1.25 for the OHCA incidence were identified in high-risk areas in the Southeast (Gangwon) and Northwest (Jeju) regions, which was consistent with findings for the SIR (Fig. 4A). Notably, a decreasing incidence was observed, particularly in major metropolitan areas centered around regional emergency medical centers. Regarding survival outcomes, areas with a higher RR, which are colored in blue in Fig. 4B to E, had better survival outcomes. Among survival outcomes, prehospital ROSC had the largest number of areas marked in blue. Based on an RR > 1.25, survival to hospital admission was high in the metropolitan areas (Seoul and part of Gyeonggi-do) and metropolitan cities (Gwangju and Daejeon). Similarly, survival to hospital discharge exhibited a higher RR in the Southwest and West Coast areas.

    Fig. 4
    EP for a specific threshold (relative risk > 1.25) of the OHCA incidence and survival outcomes. The EP was estimated from the model via spatiotemporal poisson regression. (A) Incidence of OHCA. (B) Prehospital return of spontaneous circulation. (C) Survival to hospital admission. (D) Survival to hospital discharge. (E) Good neurological recovery. Regarding incidence, risk areas are identified by red marking and survival outcomes by blue to identify areas with improved results. Regional emergency medical centers in the area marked as black dots.
    EP = exceedance probability, OHCA = out-of-hospital cardiac arrest.

    Supplementary Figs. 9 and 10 present the EP for the OHCA incidence in the 250 districts and in medical cases, respectively. When comparing the incidence of EP for the overall OHCA cases with that for medical cases, a pattern similar to that shown in Fig. 4A and Supplementary Fig. 3 emerged. For the overall OHCA, many areas revealed RR values higher than 1.25, whereas for medical cases, the RR values appeared lower than those for overall OHCA. Furthermore, consistently elevated risk levels were observed in the northeast (Gangwon) and southwest (Jeju) areas. Over time, a similar trend was observed where spatially high-risk areas gradually decreased.

    The EPs of the survival outcomes of all years are presented in Supplementary Figs. 11 to 14. The regions with RR > 1.25 were mainly located in the central and northeastern areas, along with some southeastern areas. RR ≥ 1.25 predominantly indicated favorable outcomes in major urban areas. Additionally, near the western coast, higher survival rates were observed.

    DISCUSSION

    This study analyzed spatiotemporal risks associated with the incidence and survival outcomes of OHCA at the regional level in Korea from 2009 to 2021. Despite an increase in the number of OHCA due to recent population aging, our analysis revealed a visual confirmation of a decreasing trend in the SIR for several regions and a consistent decline of the temporal trend of the RR since 2016. This study also highlights the importance of spatiotemporal epidemiology in understanding the spread of OHCA by utilizing geospatial data and statistical methods to identify high-risk areas and spatial and temporal heterogeneity. These findings further revealed that improvements in prehospital ROSC, survival to admission, and survival to hospital discharge were important outcomes. Despite the increase in the number of cases, survival outcomes improved over the 3 timepoints (2009, 2015, and 2021).

    While substantial improvements were observed in prehospital ROSC, the enhancement in the RR of survival to admission and survival to discharge was primarily evident in major urban areas. However, rural and mountainous regions continued to exhibit relatively lower risk mappings. Fig. 1B and C show that these areas have low population density and a high aging index. Furthermore, in regions outside the metropolitan areas, slower improvements were noted in terms of good neurological recovery than in other survival outcomes. Consistent with prior research, the high incidence rates in the high-risk areas may be attributable to several factors, including low rates of BCPR, prolonged dispatch and transfer times of 119 ambulance teams, and absence of emergency medical institutions in those regions.8, 41 The reasons for the disparities include the longer time required to transport EMSs, inadequate regional emergency medical centers, and small populations comprising mainly older adults. Furthermore, the survival outcomes from 2020 to 2021 declined (Fig. 2), which is consistent with prior studies suggesting a worsening of OHCA outcomes during the coronavirus disease 2019 (COVID-19) pandemic.42, 43 However, to comprehensively assess the impact of COVID-19, ongoing and systematic long-term epidemiological descriptive studies are essential. Through such studies, we can better understand whether survival outcomes will improve or challenges would continue to occur in the extended aftermath of the pandemic.

    In addition, in this study, the CPR training experience rate and AED practice experience rate were also examined. The CPR training experience rate improved by city, county, and district, but the AED practice experience rate was relatively low. Even in the metropolitan areas and major metropolitan areas, it is necessary to increase the CPR training experience rate, and the AED practice experience rate can be increased nationwide to help improve the survival rate. Therefore, it is necessary to reduce the geographical disparities to improve the survival outcomes, particularly good neurological recovery, in rural areas.31

    Studies using OHCA and spatial analysis within geographic information systems have reported risk mapping using hotspot analysis (Moran’s I, Local Moran’s I statistics) or SaTScan’s spatial scan statistics to identify high-risk areas.44, 45 However, these studies had limitations, as they did not cover several years or only studied some regions. Next, spatiotemporal analyses investigated the spatial variability of the OHCA risk and the optimal deployment of BCPR and AEDs within the chain of survival.10, 13, 14, 15, 16, 17, 20, 22, 23 Research on OHCA has primarily focused on identifying clusters and hotspots in specific regions or on investigating spatiotemporal patterns. In Korea, studies have compared urban and rural communities based on epidemiological characteristics, but these studies did not consider time and space simultaneously.31

    This study overcomes the limitations of previous studies and presents several distinct advantages. Utilizing the INLA method and Bayesian hierarchical modeling for spatiotemporal analysis was a unique approach for understanding the spatiotemporal epidemiology of OHCA in Korea. Furthermore, this study included all OHCA cases over 13 years from the initiation of the investigation until 2021; our results are from an illustrative epidemiological study and were compared with those of previous studies. An EP map for survival outcomes was presented in addition to the risk mapping of RR, which distinguishes the uniqueness of identifying risk areas based on the RR threshold. In Korea, research using EP has primarily focused on COVID-19 or environmental fields.46, 47, 48 More detailed geographic and result presentations could be made by providing an EP map explaining excess probability and presenting estimated risks of OHCA occurrences and survival outcomes, thus enabling a clearer prioritization of public health interventions within a limited environment.

    While this study offers valuable insights into the spatiotemporal epidemiology of OHCA incidence and survival outcomes in Korea, it has several limitations. First, it was a descriptive epidemiological study; other factors that could cause spatiotemporal changes were not considered. Second, patient comorbidities and individual-level covariates were not considered. Nevertheless, our findings provide significant insights into the spatiotemporal epidemiology of OHCA in Korea. Caution should be exercised in generalizing the results to other settings or drawing causal inferences from these findings. Despite these limitations, this study has the advantages of utilizing spatiotemporal interaction analysis to 1) examine the longest period (13 years) to date, 2) focus on individual-level small-area units, and 3) present both incidence and survival outcomes.

    By conducting a spatiotemporal analysis of the OHCA incidence and survival outcomes, this study provides valuable information for identifying high-risk areas and implementing targeted public health interventions. Moreover, this study highlights the importance of adopting a chain of survival-based approach to enhance survival outcomes, including good neurological recovery. By considering both temporal and spatial aspects, this study provides a more comprehensive understanding of OHCA. It facilitates the formulation of effective strategies for OHCA management and improved survival.

    In conclusion, this study provides important insights into the spatiotemporal epidemiology of the OHCA incidence and survival outcomes in Korea. These findings have practical implications for guiding targeted interventions and policies to enhance OHCA outcomes in Korea while highlighting the importance of spatiotemporal analysis in understanding and addressing public health challenges. Furthermore, the district-level analysis approach offers a valuable tool for identifying high-risk areas and populations, which can aid the development of tailored interventions to improve OHCA outcomes in these areas. Further investigation with risk factors at a district level is required to improve survival outcomes.

    SUPPLEMENTARY MATERIALS

    Supplementary Data 1

    Model summaries

    Click here to view.(200K, pdf)

    Supplementary Table 1

    Comparison of the model fit for the OHCA incidence and survival outcomes using the WAIC

    Click here to view.(126K, pdf)

    Supplementary Table 2

    Details of survival outcomes for men by age group between 2009 and 2021

    Click here to view.(120K, pdf)

    Supplementary Table 3

    Details of survival outcomes for women by age group between 2009 and 2021

    Click here to view.(115K, pdf)

    Supplementary Fig. 1

    Administrative district map of the Republic of Korea.

    Click here to view.(363K, pdf)

    Supplementary Fig. 2

    Map of risk factors associated with out-of-hospital cardiac arrest.

    Click here to view.(200K, pdf)

    Supplementary Fig. 3

    Standardized incidence ratios for cardiac causes at the district level for 2009, 2015, and 2021.

    Click here to view.(172K, pdf)

    Supplementary Fig. 4

    Posterior mean of the relative risk of prehospital return of spontaneous circulation for 250 administrative districts from 2009 to 2021.

    Click here to view.(543K, pdf)

    Supplementary Fig. 5

    Posterior mean of the relative risk of survival to hospital admission for 250 administrative districts from 2009 to 2021.

    Click here to view.(548K, pdf)

    Supplementary Fig. 6

    Posterior mean of the relative risk of survival to hospital discharge for 250 administrative districts from 2009 to 2021.

    Click here to view.(546K, pdf)

    Supplementary Fig. 7

    Posterior mean of the relative risk of good neurological recovery for 250 administrative districts from 2009 to 2021.

    Click here to view.(542K, pdf)

    Supplementary Fig. 8

    Map of risk factors associated with survival outcomes. (A) CPR education experience in 2012, 2014, 2016, 2018, and 2020. (B) AED mannequin practice experience in 2020.

    Click here to view.(231K, pdf)

    Supplementary Fig. 9

    Exceedance probability for a specific threshold (relative risk > 1.25) of the out-of-hospital cardiac arrest incidence for 250 administrative districts from 2009 to 2021.

    Click here to view.(410K, pdf)

    Supplementary Fig. 10

    Exceedance probability for a specific threshold (relative risk > 1.25) of the out-of-hospital cardiac arrest incidence for 250 administrative districts from 2009 to 2021.

    Click here to view.(238K, pdf)

    Supplementary Fig. 11

    Exceedance probability for a specific threshold (relative risk > 1.25) of prehospital return of spontaneous circulation for 250 administrative districts from 2009 to 2021.

    Click here to view.(396K, pdf)

    Supplementary Fig. 12

    Exceedance probability for a specific threshold (relative risk > 1.25) of survival to hospital admission for 250 administrative districts from 2009 to 2021.

    Click here to view.(620K, pdf)

    Supplementary Fig. 13

    Exceedance probability for a specific threshold (relative risk > 1.25) of survival to hospital discharge for 250 administrative districts from 2009 to 2021.

    Click here to view.(408K, pdf)

    Supplementary Fig. 14

    Exceedance probability for a specific threshold (relative risk > 1.25) of good neurological recovery for 250 administrative districts from 2009 to 2021.

    Click here to view.(406K, pdf)

    Notes

    Funding:This research was supported by the National Fire Agency of Korea and Korea Disease Control and Prevention Agency (grant number: 11-1790387-000878-01).

    Disclosure:The authors have no potential conflicts of interest to disclose.

    Author Contributions:

    • Conceptualization: Lee N, Jung S, Ro YS, Park JH, Hwang SS.

    • Data curation: Lee N, Jung S, Ro YS, Park JH, Hwang SS.

    • Formal analysis: Lee N, Jung S, Hwang SS.

    • Methodology: Lee N, Jung S, Ro YS, Park JH, Hwang SS.

    • Validation: Lee N, Jung S, Ro YS, Park JH, Hwang SS.

    • Visualization: Lee N, Jung S, Hwang SS.

    • Writing - original draft: Lee N.

    • Writing - review & editing: Lee N, Jung S, Ro YS, Park JH, Hwang SS.

    References

      1. McNally B, Robb R, Mehta M, Vellano K, Valderrama AL, Yoon PW, et al. 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(8):1–19.
      1. Myat A, Song KJ, Rea T. Out-of-hospital cardiac arrest: current concepts. Lancet 2018;391(10124):970–979.
      1. Mal PR, Sunnel P. Out-of-hospital cardiac arrest (OHCA): a critical healthcare problem. ECEC 2019;4:197–204.
      1. Berdowski J, Berg RA, Tijssen JG, Koster RW. Global incidences of out-of-hospital cardiac arrest and survival rates: systematic review of 67 prospective studies. Resuscitation 2010;81(11):1479–1487.
      1. Brooks SC, Clegg GR, Bray J, Deakin CD, Perkins GD, Ringh M, et al. Optimizing outcomes after out-of-hospital cardiac arrest with innovative approaches to public-access defibrillation: a scientific statement from the International Liaison Committee on Resuscitation. Circulation 2022;145(13):e776–e801.
      1. Kim HS, Chun JH. Determinants on survival outcomes of sudden out-of-hospital cardiac arrest: a multilevel analysis. Korean J Emerg Med Serv 2020;24(2):7–26.
      1. Roh SY, Choi JI, Park SH, Kim YG, Shim J, Kim JS, et al. The 10-year trend of out-of-hospital cardiac arrests: a Korean nationwide population-based study. Korean Circ J 2021;51(10):866–874.
      1. Korea Disease Control and Prevention Agency. Sudden cardiac arrest survey statistics report. [Updated 2021]. [Accessed March 20, 2023].
        https://www.kdca.go.kr/injury/biz/injury/recsroom/statsSmMain.do;jsessionid=fmr-S-ZnRqrC9czcb5QeLbmEAW_TBn1HuNfYG6AI.injury10.
      1. The Lancet. Out-of-hospital cardiac arrest: a unique medical emergency. Lancet 2018;391(10124):911.
      1. Gowens P, Smith K, Clegg G, Williams B, Nehme Z. Global variation in the incidence and outcome of emergency medical services witnessed out-of-hospital cardiac arrest: a systematic review and meta-analysis. Resuscitation 2022;175:120–132.
      1. Lerner EB, Rea TD, Bobrow BJ, Acker JE 3rd, Berg RA, Brooks SC, et al. Emergency medical service dispatch cardiopulmonary resuscitation prearrival instructions to improve survival from out-of-hospital cardiac arrest: a scientific statement from the American Heart Association. Circulation 2012;125(4):648–655.
      1. Rhee BY, Kim B, Lee YH. Effects of prehospital factors on survival of out-of-hospital cardiac arrest patients: age-dependent patterns. Int J Environ Res Public Health 2020;17(15):5481.
      1. Chen CC, Chen CW, Ho CK, Liu IC, Lin BC, Chan TC. Spatial variation and resuscitation process affecting survival after out-of-hospital cardiac arrests (OHCA). PLoS One 2015;10(12):e0144882
      1. Doan TN, Wilson D, Rashford S, Ball S, Bosley E. Spatiotemporal variation in the risk of out-of-hospital cardiac arrests in Queensland, Australia. Resusc Plus 2021;8:100166
      1. Auricchio A, Peluso S, Caputo ML, Reinhold J, Benvenuti C, Burkart R, et al. Spatio-temporal prediction model of out-of-hospital cardiac arrest: Designation of medical priorities and estimation of human resources requirement. PLoS One 2020;15(8):e0238067
      1. Peluso S, Mira A, Rue H, Tierney NJ, Benvenuti C, Cianella R, et al. A Bayesian spatiotemporal statistical analysis of out-of-hospital cardiac arrests. Biom J 2020;62(4):1105–1119.
      1. Wong PP, Low CT, Cai W, Leung KT, Lai PC. A spatiotemporal data mining study to identify high-risk neighborhoods for out-of-hospital cardiac arrest (OHCA) incidents. Sci Rep 2022;12(1):3509.
      1. Lim SL, Smith K, Dyson K, Chan SP, Earnest A, Nair R, et al. Incidence and outcomes of out-of-hospital cardiac arrest in Singapore and Victoria: a collaborative study. J Am Heart Assoc 2020;9(21):e015981
      1. 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(6):752–759.
      1. Sun CL, Karlsson L, Torp-Pedersen C, Morrison LJ, Folke F, Chan TC. Spatiotemporal AED optimization is generalizable. Resuscitation 2018;131:101–107.
      1. Baumgarten MC, Röper J, Hahnenkamp K, Thies KC. Drones delivering automated external defibrillators-Integrating unmanned aerial systems into the chain of survival: a simulation study in rural Germany. Resuscitation 2022;172:139–145.
      1. Kao JH, Chan TC, Lai F, Lin B, Sun W, Chang K, et al. Spatial analysis and data mining techniques for identifying risk factors of out-of-hospital cardiac arrest. Int J Inf Manage 2017;37(1):1528–1538.
      1. Fleming D, Owens A, Eckstein M, Sanko S. Spatiotemporal analysis of out-of-hospital cardiac arrest in the City of Los Angeles, 2011–2019. Resuscitation 2021;165:110–118.
      1. Hwang SS, Ahn KO, Shin SD, Ro YS, Lee SY, Park JO, et al. Temporal trends in out-of-hospital cardiac arrest outcomes in men and women from 2008 to 2015: a national observational study. Am J Emerg Med 2021;41:174–178.
      1. Ro YS, Hwang SS, Shin SD, Han D, Kang S, Song KJ, et al. Presumed regional incidence rate of out-of-hospital cardiac arrest in Korea. J Korean Med Sci 2015;30(10):1396–1404.
      1. Korean Statistical Information Service. Summary of census population. [Updated 2023]. [Accessed April 10, 2023].
        https://kosis.kr/statHtml/statHtml.do?orgId=101&tblId=DT_1BPA001&conn_path=I2.
      1. Song J, Ahn S, Kim J, Cho H, Moon S, Choi SH, et al. Sex-related disparities in the in-hospital management of patients with out-of-hospital cardiac arrest. Resuscitation 2022;173:47–55.
      1. Choi DH, Ro YS, Kim KH, Park JH, Jeong J, Hong KJ, et al. The association between alcohol intake shortly before arrest and survival outcomes of out-of-hospital cardiac arrest. Resuscitation 2022;173:39–46.
      1. Kwon JM, Jeon KH, Kim HM, Kim MJ, Lim S, Kim KH, et al. High incidence and mortality of out-of-hospital cardiac arrest on traditional holiday in South Korea. Korean Circ J 2019;49(10):945–956.
      1. Choi Y, Park JH, Jeong J, Kim YJ, Song KJ, Shin SD. Extracorporeal cardiopulmonary resuscitation for adult out-of-hospital cardiac arrest patients: time-dependent propensity score-sequential matching analysis from a nationwide population-based registry. Crit Care 2023;27(1):87.
      1. Ro YS, Shin SD, Song KJ, Lee EJ, Kim JY, Ahn KO, et al. A trend in epidemiology and outcomes of out-of-hospital cardiac arrest by urbanization level: a nationwide observational study from 2006 to 2010 in South Korea. Resuscitation 2013;84(5):547–557.
      1. Becher H, Winkler V. Estimating the standardized incidence ratio (SIR) with incomplete follow-up data. BMC Med Res Methodol 2017;17(1):55.
      1. Rue H, Martino S, Chopin N. Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations. J R Stat Soc Series B Stat Methodol 2009;71(2):319–392.
      1. Blangiardo M, Cameletti M, Baio G, Rue H. Spatial and spatio-temporal models with R-INLA. Spat Spatio-Temporal Epidemiol 2013;4:33–49.
      1. Naing NN. Easy way to learn standardization: direct and indirect methods. Malays J Med Sci 2000;7(1):10–15.
      1. Moraga P. Small area disease risk estimation and visualization using R. R J 2018;10(1):495.
      1. Richardson S, Thomson A, Best N, Elliott P. Interpreting posterior relative risk estimates in disease-mapping studies. Environ Health Perspect 2004;112(9):1016–1025.
      1. Rohleder S, Bozorgmehr K. Monitoring the spatiotemporal epidemiology of Covid-19 incidence and mortality: a small-area analysis in Germany. Spat Spatio-Temporal Epidemiol 2021;38:100433
      1. Jaya IG, Folmer H. Bayesian spatiotemporal forecasting and mapping of COVID-19 risk with application to West Java Province, Indonesia. J Reg Sci 2021;61(4):849–881.
      1. R Core Team. R: a language and environment for statistical computing. [Updated 2022]. [Accessed March 20, 2023].
        https://www.R-project.org/.
      1. Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from out-of-hospital cardiac arrest: a systematic review and meta-analysis. Circ Cardiovasc Qual Outcomes 2010;3(1):63–81.
      1. Bielski K, Szarpak A, Jaguszewski MJ, Kopiec T, Smereka J, Gasecka A, et al. The influence of COVID-19 on out-hospital cardiac arrest survival outcomes: an updated systematic review and meta-analysis. J Clin Med 2021;10(23):5573.
      1. Kim YS, Lee SH, Lim HJ, Hong WP. Impact of COVID-19 on out-of-hospital cardiac arrest in Korea. J Korean Med Sci 2023;38(12):e92
      1. Sasson C, Cudnik MT, Nassel A, Semple H, Magid DJ, Sayre M, et al. Identifying high-risk geographic areas for cardiac arrest using three methods for cluster analysis. Acad Emerg Med 2012;19(2):139–146.
      1. Nassel AF, Root ED, Haukoos JS, McVaney K, Colwell C, Robinson J, et al. Multiple cluster analysis for the identification of high-risk census tracts for out-of-hospital cardiac arrest (OHCA) in Denver, Colorado. Resuscitation 2014;85(12):1667–1673.
      1. Lym Y, Lym H, Kim K, Kim KJ. Spatiotemporal associations between local safety level index and COVID-19 infection risks across capital regions in South Korea. Int J Environ Res Public Health 2022;19(2):824.
      1. Yum SG, Wei HH, Jang SH. Estimation of the non-exceedance probability of extreme storm surges in South Korea using tidal-gauge data. Nat Hazards Earth Syst Sci 2021;21(8):2611–2631.
      1. Lee M, An H, Jeon S, Kim S, Jung K, Park D. Development of an analytical probabilistic model to estimate runoff event volumes in South Korea. J Hydrol (Amst) 2022;612:128129
    MeSH Terms
    Epidemiology
    Exhibition
    Incidence*
    Korea*
    Observational Study
    Odds Ratio
    Out-of-Hospital Cardiac Arrest*
    Public Health
    Resource Allocation
    Return of Spontaneous Circulation
    Small-Area Analysis
    Spatio-Temporal Analysis*
    Metrics
    Share
    Figures
    slide
    slide
    slide
    slide

    1 / 4

    Tables
    slide

    1 / 1

    ORCID IDs
    Funding Information
    PERMALINK