We analyzed country-specific epidemiological data on COVID-19 cases and international travel volume in South Korea from January to October 2020. First, data on the number of confirmed COVID-19 cases in South Korea were extracted from the Korea Disease Control and Prevention Agency (publicly available data) [18]. The epidemiological data included the dates of confirmation, dates of symptom onset, and transmission classification (local transmission/imported cases) [19]. Second, data on the monthly number of passengers entering South Korea in 2019-2020 were gathered from Incheon International Airport (publicly available data) [20]. Third, data on the number of confirmed cases of COVID-19 from the countries of origin were collected from the WHO situation report [21] and countries’ populations were obtained from [22]. The country-specific data are presented in Table S1 and Figures S1-S5 in Multimedia Appendix 1.

We have presented the epidemic curves of local and imported cases in South Korea in Figure 1A and in Seoul and Gyeonggi in Figure 1B. Most COVID-19 cases occurred in March 2020 (due to the explosive outbreaks in Daegu and Gyeongbuk), and the number of cases in Seoul and Gyeonggi increased steadily from May 2020, leading to a larger outbreak in September 2020. Figure 1B shows the epidemic curve in Seoul and Gyeonggi along with the timeline of screening and quarantine interventions (green box) and social distancing interventions (gray box). The list of selected interventions is presented in Table S2 in Multimedia Appendix 1. The COVID-19 transmission dynamics in South Korea showed spatial heterogeneity. There were two major hotspots. First, the early outbreak was primarily in the Daegu and Gyeongbuk areas from February to April 2020 due to the Shincheonji Church–related clusters, as shown in Figure 1A and reported previously [23,24]. Second, the late outbreak was primarily in the Seoul and Gyeonggi areas in September and November 2020, as it was triggered by the Sarangjeil Church–related gathering on August 15, 2020, as shown in Figure 1A.

The timeline of the administrative measures implemented in South Korea is shown in Figure 1B and Table S2 in Multimedia Appendix 1. The Korean government, Seoul, and Gyeonggi implemented administrative countermeasures in response to the COVID-19 outbreak, including guidelines for entry restrictions followed by the 2-week self-quarantine guidelines from a different period combined with social distancing interventions.

We aimed to calculate the country-specific importation risk of COVID-19 based on the number of international travelers, confirmed cases in the originating countries, and the population of the originating countries. The countries were grouped as Europe (the United Kingdom, Germany, and France); China and Asia except China; and North America (the United States and Canada). Country-specific importation risk is defined as a function of three factors: population, the number of confirmed cases of COVID-19, and passengers entering South Korea [25,26]. The risk of importation from a given country in a given month was derived as follows:

In this section, a renewal equation was employed to estimate the R₀. The R₀ is defined as the average number of susceptible individuals infected by a single primary case. Previous studies on COVID-19 estimated the R₀ to be 2-3 [26-28]. In this study, we have categorized the total cases into locally transmitted (local cases) and imported cases. The total incidence of COVID-19 at time t, denoted by i(t), is the sum of local cases (i_L(t)) and imported cases (i_o(t))—that is, i(t)=i_o(t)+i_L(t). The renewal equation for the transmission dynamics of COVID-19 is defined as follows [29,30]:

The data are presented in Table S1 in Multimedia Appendix 1. The data sets were fully anonymized and did not include any personally identifiable information. Thus, ethical approval was not required for the analysis.

Relation between the number of passengers and imported cases was explored and the normalized country-specific risk was obtained. Among four groups of countries, China had a high risk of importation until February 2020. Afterward, North America showed a high risk of importation. The number of imported cases in Korea had a high correlation with the normalized risk, and the Spearman correlation coefficient was 0.82. The R₀ was estimated at 1.87 (95% CI 1.47-2.34) with the rate of α=0.07 in Seoul and Gyeonggi. R₀ was varied according to α to be between 1.83 and 3.94 in South Korea. The R_t in South Korea and in Seoul and Gyeonggi were shown and interpreted along with the control interventions.

Figure S1 in Multimedia Appendix 1 illustrates a summary of the epidemiological data on COVID-19 cases in South Korea. The top panels of Figure S1A in Multimedia Appendix 1 show the overall characteristics of the confirmed cases from February to October 2020. The leftmost panel shows the ratio of COVID-19 cases by region: 43% in the Seoul and Gyeonggi areas, 33% in the Daegu and Gyeongbuk areas, and 24% in the rest of South Korea. The second panel shows the ratio of imported cases (14%) to local cases (86%). The third panel shows the number of confirmed cases in Seoul and Gyeonggi and the ratio of imported cases (10%) to local cases (90%). The rightmost panel shows that 31% of imported cases were reported in Seoul and Gyeonggi. Seoul and Gyeonggi are the regions with the most inflow of foreigners in South Korea as they have major international airports (ie, Incheon International Airport and Gimpo International Airport).

Figure S1B in Multimedia Appendix 1 shows the monthly proportion of imported cases from five continents and the total proportion of imported cases (top left panel). A list of country-specific imported cases is given in Table S1 in Multimedia Appendix 1. The number of imported cases from China was large during January and April 2020, while that from North America and Europe increased until April 2020. However, the number of imported cases from Asia increased rapidly from May to October 2020; Asia (50.6%), North America (28.1%), and Europe (17.8%) accounted for most of the cumulative imported cases.

The importation risk implies that a country with more COVID-19 cases and more travelers entering South Korea has a higher risk of importation. Figure 2 shows the relationship between the number of passengers entering South Korea in 2020 and the monthly confirmed cases of COVID-19 between January and April in 2020. After February 2020, the number of Chinese passengers rapidly decreased due to the emerging outbreak of COVID-19 in China, as shown in Figures 2A and 2B. The number of confirmed cases and the number of passengers entering South Korea from the United States increased from March to October 2020, as shown in Figure S2 in Multimedia Appendix 1. This indicated that the number of passengers was dramatically reduced since the COVID-19 pandemic began, owing to travel bans and restrictions (see Table S2 in Multimedia Appendix 1). Next, the number of international travelers who arrived in South Korea in 2020 was compared with that in 2019 (Figure S3 in Multimedia Appendix 1). The number of country-specific confirmed cases and passengers per month from the top 13 countries from January to October 2020 are shown in Figures S4 and S5 in Multimedia Appendix 1, respectively. The number of cases in China was reduced dramatically from March 2020 (Figure S4A in Multimedia Appendix 1), while the number of cases in the United States, the United Kingdom, France, Germany, Canada, and Malaysia continued to increase until October 2020 (Figures S4H-M in Multimedia Appendix 1). However, the number of passengers was greatly reduced, regardless of country, as shown in Figure S5 in Multimedia Appendix 1.

We presented the risk of country-specific importation in Figure 3 and Table 1. Between January and February 2020, the risk of importation from China was the highest among the countries studied, while that from North America (the United States and Canada) showed a significantly high risk of importation from April to October 2020. The number of imported cases was highly correlated with the normalized risk (the Spearman correlation coefficient and Kendall correlation coefficient were 0.82 and 0.64, respectively).

Figure 4 showed the results of regression analysis between the country-specific risk and the number of imported cases according to the four different countries/regions. The estimated values using the regression analysis are summarized in Table S3 in Multimedia Appendix 1. All countries had a positive relationship between imported cases and the risk of importation because the estimates of the slopes were positive. The linear regression models were well fitted, especially for China (R²=0.78). Asia (except China) was the most affected region with respect to the risk of importation. It is clear that imported cases entering from Asia (except China) can increase much more if the risk importation of Asia (except China) is elevated.

Reproduction Number

We varied α from 0 to 1 and estimated the R₀ for the early COVID-19 outbreak in Seoul and Gyeonggi based on COVID-19 confirmed cases from January 10 to February 25, 2020, while the R₀ in South Korea was based on COVID-19 confirmed cases from February 1 to February 19 (Table 2). Figure 5 shows the comparison between the COVID-19 data and estimated cases (Figure 5A) and cumulative local cases (Figure 5B). The corrected Akaike information criterion was calculated at 159.08 and the Bayesian information criterion was 160.84. Table 2 illustrates the estimation of R₀ with varying α. The estimated R₀ was 1.87 (95% CI 1.47-2.34) with α=0.07 and 1.49 (95% CI 1.17-1.87) with α=1.0. This indicated that the value of R₀ decreased with increasing α since there was no secondary infection from imported cases (α=1).

The R_t was calculated for α=0.07 and the daily epidemic curves of imported (green) and local (gray) cases and the R_t values (blue curve) in the Seoul and Gyeonggi regions are shown in Figure 6. On March 11, 2020, when the WHO declared a pandemic situation, the R_t value fell below 1. At the beginning of the holiday season in early May 2020, the value again increased rapidly. In addition, there was a large outbreak in August and September 2020 due to a public gathering from all over South Korea on August 15, 2020. We conducted the sensitivity analysis for R_t according to the different time windows and α values (Figures S6A and S6B in Multimedia Appendix 1). Additionally, the R_t in South Korea from April to October 2020 is shown in Figures S6C and S6D in Multimedia Appendix 1.

In this study, we analyzed country-specific epidemiological data and data on passengers entering South Korea from January to October 2020. First, the correlation between the number of passengers and COVID-19 cases by country was calculated. A country with more confirmed cases showed a higher risk of importation. Second, the country-specific risk was highly correlated with the number of imported cases by country (Spearman correlation coefficient=0.82); China had the highest importation risk of COVID-19 in the early stages of the pandemic (January and February 2020), while North America (United States and Canada) showed the high importation risk from April to October 2020. Third, for the early stages of the COVID-19 pandemic, the R₀ was estimated at 1.87 (95% CI 1.47-2.34), which was similar to the R₀ of COVID-19 of approximately 2-3 in Wuhan, China [28,35,36]. Finally, we estimated the R_t by employing the renewal equation, accounting for the effects of control interventions.

Mainland China and South Korea experienced a steep rise in the number of COVID-19 confirmed cases in the early stages of the COVID-19 pandemic [3,32]. It appears that the two governments’ respective reactions to the novel virus resulted in a successful reduction in infection rates; however, the cost burden that the two countries had to bear was quite different. China implemented a lockdown in Hubei province and strict border control measures against higher-risk countries [37,38]. In China, active measures driven by the central government to retard the progress of epidemic diseases appear to be effective in impeding the spread of COVID-19; however, Chinese people had to pay burdensome costs during the initial outbreak of COVID-19 [39]. South Korea used strict social distancing measures and self-quarantine without restricting borders. However, South Korea expanded the volume of testing and promptly traced the contacts of confirmed cases [32,40].

There were potential risk factors that could have led to a larger outbreak of COVID-19 in South Korea. First, negative serial intervals indicated presymptomatic transmissions, highlighting the potential risk of transmission from asymptomatic cases. In South Korea, 12.7% of serial intervals were negative (199 pairs among 1567 symptomatic pairs), which could lead to large outbreaks as secondary transmission before the appearance of symptoms cannot be prevented [40,41]. Second, the R₀ was estimated at 1.49 in Seoul and Gyeonggi, indicating that an epidemic might be possible. Third, the number of imported cases increased substantially from April 2020, although passenger volume has been rapidly decreasing due to the increase in COVID-19 cases. However, a large outbreak of COVID-19 caused by imported cases did not occur because of the policy of testing passengers arriving from other countries and isolating them for a minimum of 2 weeks; this policy was implemented on April 1, 2020, and reduced the risk of the spread of COVID-19.

This study has several limitations. First, this study relied on confirmed cases in South Korea. However, there were a substantial number of asymptomatic infections as 12.5% of serial intervals were negative [41], which represents presymptomatic transmission [32,40]. Thus, we did not consider secondary transmission caused by imported and local cases. Second, we analyzed the impact of imported cases on the local transmission of COVID-19 and found that the impact of imported cases had decreased since the strict implementation of airport screening and quarantine measures from April 2020 [24,42] (Table S2 in Multimedia Appendix 1). Therefore, the potential risk of infection from imported cases was regarded as a less important factor. However, the reason for the small outbreaks resulting from imported cases was due to strengthened screening inspection and self-quarantine measures for those entering South Korea, which were implemented on April 1, 2020. This means that secondary transmission by imported cases can play a critical role in COVID-19 transmission. Finally, the risk was estimated by month as monthly passenger volume data were available. If daily data were given, the risk by country could be computed daily or weekly. However, we observed a significantly different risk of importation of COVID-19 from overseas countries.

Despite these limitations, we investigated the risk of importation of COVID-19 using country-specific epidemiological data and passenger volume. By combining social distancing, screening, and self-quarantine for all travelers entering South Korea, the mitigation of COVID-19 transmission caused by imported cases in South Korea was highly successful. These efforts—accompanied by identification of the source of infection and strengthened quarantine measures for travelers from overseas countries—should be continued. Therefore, it is urgent to assess the risk of importation and maintain an effective surveillance system for COVID-19 in South Korea.

Strict control interventions were implemented to prevent the spread of COVID-19 in South Korea since the first case was confirmed in the country on January 20, 2020. The COVID-19 outbreak in South Korea has been successfully suppressed without strict lockdowns. First, the Korean government has constructed a rapid testing and diagnosis system [40,43]. Previous studies have shown that most cases have been confirmed within a week of illness onset [24,43]. Moreover, drive-through screening centers were initiated on February 23, 2020, in Daegu, South Korea [44]. This system contributed to the rapid diagnosis and further testing of suspected cases. The entire drive-through testing procedure takes about 10 minutes, and it is helpful for diagnosing infections early in cases with mild symptoms or asymptomatic cases. Second, the widespread epidemiological investigation of contact tracing was conducted in infected as well as suspected cases [40]. Social distancing strategies and mask wearing have been recommended since February 2020. Social distancing has been found to mitigate the spread of COVID-19 cases [40,45,46]. Third, the Korean government introduced a “special entry procedure,” which was applied to all passengers from mainland China from February 4, 2020, to control imported cases. Subsequently, all passengers from overseas countries were quarantined for 14 days after April 1, 2020. Combined control interventions, including social distancing efforts, appear to have succeeded in preventing the spread of COVID-19 in South Korea.

Since the severity of COVID-19 outbreaks and health policies differ across countries, a country-specific surveillance system would be more efficient than a uniform screening and surveillance policy for every country. Another notable feature of South Korean border control is that the number of international travel hubs is limited, and passenger traffic can be effectively monitored. This, in turn, helps in diagnosing and tracing imported cases without imposing a strict lockdown. It is important to estimate the country-specific risk of importation by identifying high-risk countries to prevent recurrent outbreaks due to importation of cases. It would be helpful to access a finer level of information to estimate the effective reproduction number as the risk indicator of importation and local transmission. Therefore, our framework can also be applied to countries that have similar immigration policies. Furthermore, a risk assessment of imported cases between neighboring countries that do not implement border control (eg, the Schengen zone or Central America-4 Free Mobility Agreement) might be challenging.

Data on international passengers entering South Korea, the severity of the COVID-19 outbreak in originating countries, and country-specific imported cases were analyzed to compute the risk of importation of COVID-19 into South Korea. China was a high-risk country for importation in the early stages of the pandemic until March 2020, while the United States and Canada showed a high risk of importation after April 2020. Moreover, statistical model accounting was employed to estimate the R₀ and R_t using epidemiological data on imported and locally transmitted cases. Our results highlighted that rapid diagnosis and prompt implementation of case isolation and quarantine were effective in preventing secondary infections caused by imported cases through the continuous inflow of passengers traveling from high-risk countries. Therefore, multiple mitigation interventions (social distancing, a rapid diagnosis system, and movement restriction) should be implemented to reduce the spread of local and imported cases in South Korea.