Analysis of characteristics and trends of the In-Hospital Cardiac Arrest incidence in the Republic of Korea: clinical usefulness of national health insurance claim data

1. Introduction

Cardiac arrests are emergency situations that require medical professionals to provide immediate, adequate care within minutes, and cardiac arrest treatment, that is, cardio-pulmonary resuscitation (CPR), involves many human and physical resources [1, 2]. The lack of an immediate response in the event of a cardiac arrest may lead to death or permanent brain damage, and use up significant medical resources.

Cardiac arrest is traditionally classified into In-Hospital Cardiac Arrest (IHCA) and Out of Hospital Cardiac Arrest (OHCA) depending on whether or not it has occurred in a hospital [3]. For decades, various studies on the incidence and characteristics of OHCA, as well as the development of guidelines to improve the prognosis, have been actively conducted [4, 5, 6]. IHCA has tended to be considered a branch of OHCA or to be valued with low importance [7]. However, IHCA requires the preparation of considerable medical resources in hospitals [1]. Furthermore, compared to OHCA, until recently, there have not been many studies on the incidence, characteristics, and prognosis of IHCA [8].

Recently, as several registry-based studies have been conducted, the incidence rate and the characteristics of IHCA have been analyzed [8, 9]. Previous studies have shown that IHCA has approximately 1-5 occurrences per 1000 beds [2]. According to the data of the American Heart Association’s nationally representative Get With The Guidelines Resuscitation (GWTG-R) registry it was estimated that there were approximately 6-7 cases per 1000 hospitalizations (0.92/1000 bed-days) from 2003 to 2007 [10, 11]. The United Kingdom (UK) national cardiac arrest audit reported 1.6 cases per 1000 hospitalizations [12]. Denmark’s nationwide IHCA registry base incidence reported 1.8 cases per 1000 hospitalizations [9].

However, the incidence of IHCA based on real data rather than estimated data is difficult to predict [2]. Although it has been estimated based on the registry, there are not many studies have suggested the exact incidence of IHCA. The Republic of Korea is one of the relatively few countries in the world that implements national health insurance for the whole nation, and data for insurance claims can be identified on an actual scale [13]. From this insurance claim data, Choi et al. [14] and Kwon et al. [15] reported a nationwide IHCA incidence of 2.46 for every 1,000 hospitalizations using 2009 sample data and 3.00 per 1,000 hospitalizations using whole data from 2011 to 2015, respectively. Based on this pilot study, the present study attempts to grasp the IHCA incidence and analyze its characteristics by using the total data for several years.

2. Materials and methods

3. Results

3.1 Analysis based on gender and age

3.1.1 Analysis based on age

A bimodal distribution is observed when the IHCA incidence is analyzed for patients below one year old and by age group in 10-year increments. The IHCA incidence is 1.54 cases for every 1,000 hospitalizations in patients below one year old, and 1.01 cases for every 1,000 hospitalizations for 41 years onwards. The event rate is highest in the 81-90 age group (Fig. 1, Supplement Table 1).

Fig. 1.In-Hospital Cardiac Arrest Incidence Trends by Age Group. * Incidence unit: cases per 1,000 hospitalization.

The overall IHCA incidence from 2011 to 2015 is 3.00 cases per 1,000 hospitalization, and the IHCA incidences was increased in 61 years onward. However, the IHCA incidence showed a declining trend from 2011 to 2015 for patients aged 51 and older.

3.1.2 Analysis based on gender and age

The IHCA incidence of male and female patients by age group and year are shown in Fig. 2, Fig. 3, and Supplement Table 1.

Fig. 2.In-Hospital Cardiac Arrest Incidence Trends by Male and Age Group. * Incidence unit: cases per 1,000 hospitalization.

Fig. 3.In-Hospital Cardiac Arrest Incidence Trends by Female and Age Group. * Incidence unit: cases per 1,000 hospitalization.

The overall 5-year IHCA incidence was 3.92 cases per 1,000 hospitalizations for male patients, and 2.19 cases per 1,000 hospitalizations for female patients. That is, male patients had a roughly 1.8 times higher IHCA event rate.

The IHCA incidence by age group exhibited a similar binodal distribution for both male and female patients.

3.3 Analysis based on major cause of death and infectious diseases

A separate analysis was performed for major diseases - cardiovascular diseases, cerebrovascular diseases, respiratory diseases, and oncologic disease and infectious diseases (Fig. 4).

Fig. 4.In-Hospital Cardiac Arrest Incidence Trends by Major cause of Death and Infectious Disease. * Incidence unit: cases per 1,000 hospitalization.

The IHCA incidence among patients with diseases of the circulatory system, including cardiovascular diseases, showed an increasing trend at 9.57 cases per 1,000 hospitalizations in 2011, 10.41 cases in 2012, 10.36 in 2013, 11.20 in 2014, and 12.00 in 2015. The number of male patients with diseases of the circulatory system was about 1.67 times higher than the number of female patients in the five-year period, and the IHCA incidence increased for both genders.

The IHCA incidence among patients with diseases of the respiratory system, including types of pneumonia, showed a decreasing trend at 4.09 cases per 1,000 hospitalizations in 2011, 4.15 cases in 2012, 4.09 in 2013, 3.50 in 2014, and 3.42 in 2015. The number of male patients with diseases of the respiratory system was relatively higher than the number of female patients in the five-year period, and the IHCA incidence decreased for both genders.

The IHCA incidence among patients with infectious diseases decreased from 5.00 cases per 1,000 hospitalizations in 2011, 4.79 cases in 2012, 4.33 cases in 2013, and 3.85 cases in 2014 to 3.48 cases in 2015. Diseases of the nervous system including cerebrovascular diseases showed a similar decreasing trend from 1.45 cases per 1,000 hospitalizations in 2011, 1.29 cases in 2012, 1.09 cases in 2013, and 1.04 cases in 2014 to 0.86 cases in 2015.

The IHCA incidence among oncologic patients was similar each year.

3.4 Analysis based on institution type

The IHCA incidence by year and hospital type is shown in Fig. 5 and Supplement Table 3.

Fig. 5.In-Hospital Cardiac Arrest Incidence Trends by Hospital Type. * Incidence unit: cases per 1,000 hospitalization.

The IHCA incidence in tertiary hospitals showed an increasing trend at 5.43 cases per 1,000 hospitalizations in 2011, 6.25 cases in 2012, 6.13 cases in 2013, 6.31 cases in 2014, and 6.29 cases in 2015.

The IHCA incidence was generally higher for general hospitals than for tertiary hospitals, with a similar increasing trend at 5.71 cases per 1,000 hospitalizations in 2011, 6.51 cases in 2012, 6.56 cases in 2013, 6.44 cases in 2014, and 6.65 cases in 2015.

On the other hand, the IHCA event rate for hospitals and long-term care hospitals was 1.38 cases and 0.77 cases, respectively, in 2011, and dropped to 0.98 cases and 0.40 cases in 2015.

The IHCA incidence for clinic-level medical institutions increased slightly from 0.05 cases per 1,000 hospitalizations in 2011 to 0.10 cases in 2015.

3.5 Analysis based on province

Compared to the overall IHCA incidence of 3.00 cases per 1,000 hospitalizations from 2011 to 2015, the cities or provinces with a sub-overall incidence were Busan Metropolitan City, Daejeon Metropolitan City, Ulsan Metropolitan City, Jeollabuk Province, Jeollanam Province, Gyeongsangbuk Province, and Gyeongsangnam Province (Fig. 6).

Fig. 6.In-Hospital Cardiac Arrest Incidence Trends by Region. * Incidence unit: cases per 1,000 hospitalization.

Jeju Special Self-Governing Province had the highest incidence, followed by Gangwon Province, Seoul, and Chungcheongbuk Province.

The incidence of IHCA was slightly higher than the overall in three metropolitan areas, including Seoul, Incheon Metropolitan City, and Gyeonggi Province, where patients from all over the country are concentrated.

The IHCA incidence increased for Gangwon Province, Chungcheongbuk Province, and Sejong Special Self-Governing City.

Meanwhile, the IHCA event rate decreased yearly in Busan Metropolitan City, Gwangju Metropolitan City, Jeollanam Province, and Jeollabuk Province. Other provinces had similar IHCA event rates each year.

4. Discussion

In this study, the overall IHCA incidence from 2011 to 2015 was 3.00 cases for every 1,000 hospitalizations. This is consistent with the 2007 findings of Sandroni et al. [2] and Nolan et al. [16], who reported one to five IHCA cases for every 1,000 hospitalizations. In a 2017 study by Thompson et al. [17] a similar IHCA incidence of four cases per 1,000 hospitalizations was observed for U.S. elderly population aged 65 and above. Harrison et al. [12] of the United Kingdom reported 1.5 cases in acute care hospitals, while Radeschi et al. [18] of Italy observed 1.51 cases. The IHCA incidences of the two countries were slightly lower than that of Korea. China’s Shao et al. [19] declared that there were 17.5 cases for every 1,000 hospitalizations, and this was significantly higher than other countries. They also reported a low survival rate and poor prognosis. In the United States, based on the American Heart Association’s Get With The Guidelines-Resuscitation (GWTG-R) registry from 2008 to 2017, the incidence was estimated as 292,000 annually, or 9 to 10 IHCA per 1,000 admissions [8]. This difference between studies is speculated to be due to the method of calculation and the characteristics of the population.

There are few countries in the world that provide national health insurance for individual citizens. The Republic of Korea has a special health insurance system that allows individuals to make insurance claims for medical services based on their resident registration number. The insurance claims contain many types of information but may be a distorted representation as they are for insurance purposes. Moreover, the data are too vast to apply in clinical analysis. HIRA-NPS-2009, extracted from national health insurance claims in 2009 to be nationally representative, was utilized in some studies. Choi et al. [14] conducted a pilot study in 2009 using the data and found that the IHCA event rate was 2.46 cases for every 1,000 hospitalizations.

However, since most countries cannot have access to data on citizens’ medical records, IHCA incidences were only reported for certain age groups or through registries of some hospitals.

The IHCA incidence according to age group showed binodal distribution. Patients aged below one year old suffered from various diseases, from which we could predict a high IHCA incidence. Those aged 60 and above showed an increasing trend from 2011 to 2015. For patients aged 60 and above, the IHCA incidence increased with age. Nonetheless, an encouraging sign was that the IHCA event rate was decreasing annually in the same age group. This can be associated with improvements in the quality of medical services and patient safety [20].

The IHCA incidence was 1.8 times higher for male patients than female patients. This trend was also observed for patients with diseases of the circulation system and cancer, with male patients having a higher rate by 1.67 and 2.17 times, respectively. The results cannot be generalized to all diseases as the difference was less significant for diseases of the nervous system. Also, there were some conditions unique to female patients, such as pregnancy, childbirth, and puerperal infections [12, 20, 21].

Historically, the most common and second most common cause of cardiac arrest have been cardiac and respiratory origins [22, 23, 24]. In this study, the high incidence in patients with diseases classified under R00-R99, that is, “symptoms, signs and abnormal clinical and laboratory findings; not elsewhere classified”, is presumed to be the result of cardiac arrest occurring while trying to find the cause of symptoms. This result is thought to be because the data originated from the insurance claim data. Excluding IHCA related to neonatal period, cardiac disease, infectious disease, and respiratory disease are the next most common causes of IHCA, consistent with results from other studies.

Focusing on the several diseases that are regarded as to the leading causes of death, the IHCA rate of cardiovascular disease continues to increase, as opposed to be the decreasing trend of the IHCA incidence in other diseases, particularly infectious diseases, respiratory diseases (including pneumonia), and nervous system diseases. Even though this data is considered for insurance claim, the trends are thought to reflect the medical development of successful acute treatment of these diseases, and hospitals must exert efforts to lower cardiac arrest incidence among patients with diseases of the circulation system.

Looking at the IHCA incidence by hospital type, hospitals or long-term care hospitals where milder cases or patients requiring long-term care are hospitalized showed a decrease in the IHCA incidence, whereas tertiary hospitals displayed the opposite trend. This can be traced to the higher concentration of severe cases in general or tertiary hospitals. Harrison et al. [12] reported 1.5 cases for every 1,000 hospitalizations in acute care hospitals, and Radeschi et al. [18] found 1.51 for every 1,000 hospitalizations. The IHCA event rate in Korea’s tertiary hospitals and general hospitals was about four times higher at 6.07 cases and 6.37 cases, respectively. One factor that may be relevant here is the practice of CPR on patients suffering from cancer or other diseases and who are at the end of life. To resolve this issue, the Act on Hospice and Palliative Care and Decisions on Life-sustaining Treatment for Patients at the End of Life was enacted, but further research should be conducted in this area. The overdependence on CPR alone cannot explain Korea having an almost four times higher event rate compared to developed countries, which highlights the need for in-depth analysis and countermeasures.

The provinces that showed a higher IHCA incidence than the five-year national overall were Jeju Special Self-Governing Province, Gangwon Province, and Seoul. In the case of Seoul, the high incidence can be explained by the city having major hospitals preferred by many citizens. However, the high incidence cannot be explained altogether because Jeju Special Self-Governing Province is due to the isolated environment of an island, and Gangwon Province is due to the environment with many mountainous areas. Based on data of the AHA’s GWTG-R, Merchant et al. reported no statistically significant difference between incidence of IHCA in urban and rural areas [25]. This difference is due to the medical accessibility resulting from the geographical factors and relatively unforced health care delivery system in the nation [26].

Further research should be carried out to analyze the possibility of patients receiving treatment in other regions or moving to new regions, and to calibrate for severity of conditions.

While many cases of IHCA is irresistible, there are some cases that can be predicted and prevented. To prevent IHCA and respond to emergency situations, many hospitals have operated the Rapid Response Team (RRT) or Medical Emergency Team (MET) [27], and some researchers have even used tools of the deep learning, as a type of machine learning [28]. And their efforts make patient safety to be ensured by preventing cardiac arrest in hospitals. The results of this study would be expected to be utilized as a basic statistical data for their efforts.

This study has the following limitations.

First, insurance claims data contain information on examinations, treatment, procedures, surgeries, and medication, but do not provide results for each administration. As such, the cause of cardiac arrest could not be fully investigated.

Second, since not all patients’ medical records can be accessed, it was not possible to analyze the status of patients at the time of cardiac arrest.

Third, patient status in insurance claims data is briefly described as “ongoing”, “transferred”, “returned”, “death”, “other”, or “discharged”, and does not provide details on specific conditions of transfer, return, or discharge. Analysis of cerebral performance category (CPC) scores, needed to determine the survival rate or prognosis evaluation of patients who have experienced cardiac arrest, could not be carried out. To assess survival rate and prognosis, it is necessary to analyze related factors based on operational definitions. This will allow a comparative analysis of survival rate and prognosis of cardiac arrest patients between hospitals that provide CPR training on a regular basis and those that do not.

5. Conclusions

The overall IHCA incidence for patients hospitalized in medical institutions of the Republic of Korea from 2011 to 2015 was found to be 3.00 cases per 1,000 hospitalizations. The IHCA incidence was also obtained by age group, gender, disease group, institution type, and regional local government. The results of this study can be used as important basic data to secure patient safety by reducing the occurrence of IHCA.

Author contributions

IHK contributed to conduct the data curation and formal analysis and funding acquisition. YC and IHK contributed to application of the research methodology. IHK drafted the manuscript as first and corresponding authors. YC reviewed and edited for completion of the manuscript. IHK visualized the figures and tables.

Ethics approval and consent to participate

This study was approved as exempt from research ethics review with approval number DAUHIRB-EXP-17-110 issued by the Institutional Review Board of Dong-A University Hospital.

Acknowledgment

This paper is based on the In Ho Kwon’s PhD Dissertation at Kangwon National University, Gangwon, Republic of Korea. And this study used HIRA research data made by Health Insurance Review & Assessment Service. The views expressed are those of the authors and not necessarily those of HIRA and the Ministry of Health and Welfare of the Republic of Korea.

Funding

This work was supported by the Dong-A University research fund.

Conflict of interest

No potential conflict of interest relevant to this article was reported.

Supplementary material

Supplementary material associated with this article can be found, in the online version, at https://oss.signavitae.com/mre-signavitae/article/1372074403372646400/attachment/SV2021012701_Supplements.pdf.