In Korea, the National Emergency Department Information System (NEDIS) collects clinical and administrative information from approximately 400 emergency medical centers nationwide. Each year it reports national trends for 28 severe illness diagnosis codes including sepsis, but these counts are based on a narrow set of explicit codes in the Korean Classification of Diseases (KCD), and episodes labeled improperly may be missed [7].

In this study, we analyzed diagnosis codes from the NEDIS database from 2018 to 2022. Three hierarchical case definitions were applied: (1) explicit sepsis codes representing a primary or secondary KCD-8 diagnosis of sepsis or septic shock; (2) severe infection proxy denoted with any infection code [8] and either admission directly from the emergency department (ED) to an intensive care unit (ICU) or death in the ED; and (3) combination of (1) and (2). Because NEDIS does not include laboratory results or claims for specific organ-support therapies, ICU admission and ED death were used as pragmatic proxies for critical illness.

Of the 40,079,939 total patient encounters in the NEDIS database, 10,661,901 individuals (26.6%) were identified with an infection-related diagnosis, showing a mortality rate of 2.3%. Among this cohort, a smaller group of 254,690 patients (6.4 cases per 1,000 ED visits) had an explicit diagnosis code for either sepsis or septic shock (Table 1). Using a proxy definition for severe infection—any infection diagnosis combined with ICU admission or death—the encounter rate was 12.0 cases per 1,000 ED visits with a mortality of 23.8%. When this proxy was combined with sepsis and septic shock diagnosis codes for the most inclusive estimate, the rate increased to 15.4 cases per 1,000 ED visits, with an overall mortality rate of 22.8%. In the adult population (≥18 years), this comprehensive definition identified 582,244 patients, with a mortality rate of 24.1% and an encounter rate of 18.2 per 1,000 adult ED visits. In contrast, the pediatric population (<18 years) accounted for 34,448 patients, with a mortality rate of 2.1% and an encounter rate of 4.7 per 1,000 pediatric ED visits. Although this approach captures more suspected sepsis cases, it misses patients admitted to general wards without ICU care or those transferred to another hospital; applying a broader case definition with richer clinical data would likely reveal a substantially larger sepsis population.

The ED is the frontline of sepsis care, receiving 50% to 70% of all hospitalized sepsis cases [8,9]. Yet most Korean ED data are collected from university hospital–based registries or administrative reports that count only encounters carrying explicit sepsis diagnosis codes [10-12]. Consequently, the burden of sepsis, timeliness of care, and outcomes in regional and community hospitals remain largely unknown. A more comprehensive, population-wide surveillance system, either by augmenting NEDIS or by establishing a dedicated national registry, will be essential for capturing the true epidemiological burden of sepsis across all healthcare settings and for developing effective, data-driven strategies to improve patient outcomes.