Objective:Emergency department (ED) physicians face substantial cognitive and physical demands, yet workload data applicable to real-world staffing and operational decisions remain limited. This study aimed to quantify perceived workload across diverse ED tasks using the NASA Task Load Index (NASA-TLX) and to determine how workload varies by physician experience, patient acuity, and clinical context. A secondary aim was to generate practical insights that may inform resource allocation and experience-based task distribution in the ED. Methods:We conducted an observational survey of interns, residents, and specialists working in the ED of a tertiary hospital between June and July 2022. NASA-TLX questionnaires were administered to assess workload across common procedures and patient-care tasks. Analyses were stratified by physician experience, Korean Triage and Acuity Scale (KTAS) level, and chief complaint. Nonparametric methods were used to evaluate differences in workload patterns. Results:Sixty physicians participated (30 interns, 30 residents/specialists). Procedures with high technical complexity, such as thoracentesis and lumbar puncture, showed the highest workload among interns. Among residents, workload decreased from postgraduate year 1 to 3 but rose again in year 4, reflecting increased supervisory responsibilities. Higher patient acuity (KTAS 1-2) and neurological chief complaints were consistently associated with elevated workload across all experience levels. Conclusion:Perceived workload in the ED varies significantly by task type, experience level, and patient acuity. These findings provide actionable data that may support evidence-based staffing decisions, workload redistribution, and training strategies to optimize physician performance and mitigate cognitive overload in resource-limited emergency departments.

Identifying high-risk patients in emergency departments (EDs) is crucial due to high mortality rate associated with sepsis and for the timely management. This study aimed to conduct unsupervised consensus clustering of patients with suspected sepsis from a heterogeneous ED population. This descriptive, retrospective cohort study was conducted between November 2014 and November 2019 on patients with suspected sepsis who visited the ED. Cluster analysis was performed using variables, such as vital signs and clinical laboratory data, to classify patient characteristics. An artificial intelligence model generated a clustering plot based on the distance values of various variables, identifying five distinct clusters, each representing a unique clinical phenotype among the patient subtypes. In addition to analyzing the clinical phenotypes of each cluster, this study examined the prognosis based on antibiotic administration time through subgroup analysis to provide guidance for clinical application. A total of 14,402 patients were included in this study. Cluster analysis combined with an artificial intelligence model identified five distinct clusters among patients with suspected sepsis. The baseline characteristics and clinical outcomes of each of the five clusters are described in detail. Cluster B comprised the largest proportion of patients with septic shock and exhibited the highest percentage of critical care interventions, including vasopressor use, mechanical ventilation, intensive care unit admission, and 28-d mortality. Cluster E had the lowest rates of these interventions. In multivariable analysis, administration of antibiotics within 3 h significantly decreased 28-d mortality in Cluster A (aOR 0.73; 95

Bacteremia requires early detection and appropriate treatment. High rates of false-positives in blood cultures in emergency departments (EDs) have led to the need for accurate predictive tools. This study developed and validated a bacteremia prediction score to improve patient selection for blood cultures. A multicenter retrospective cohort study used data from a tertiary academic hospital in Republic of Korea (2014–2019) to develop and validate the model, with external validation at two additional hospitals (Hospital A and Hospital B) (2022). The AutoScore method, integrating machine learning and logistic regression, generated the score using initial vital signs and laboratory data. Model performance was assessed via the area under the receiver operating characteristic (AUROC) curve. Of 24,856 patients, 17,373 formed the development cohort, 2491 the validation cohort, and 4992 the test cohort. Bacteremia incidence was 11.39%. The score, incorporating eight variables (procalcitonin, lactate, white blood cell count, c-reactive protein, blood urea nitrogen, body temperature, platelet count, and total bilirubin), achieved an AUROC of 0.825 (95% confidence interval [CI], 0.808–0.842), outperforming procalcitonin (0.808; 95% CI, 0.790–0.826), SOFA score (0.718; 95% CI, 0.696–0.741), MEWS score (0.630; 95% CI, 0.606–0.654) and SIRS score (0.583; 95% CI, 0.558–0.608). External validation yielded AUROCs of 0.796 (95% CI, 0.773–0.818) at Hospital A and 0.813 (95% CI, 0.778–0.848) at Hospital B. Risk stratification categorized patients into low-, intermediate-, and high-risk groups with bacteremia rates of 2.6–3.2%, 5.2–9.3%, and 23.7–30.2%. A bacteremia prediction score was developed and validated using AutoScore, integrating machine learning and logistic regression. The model, based on eight clinical variables, stratified patients by risk and showed favorable predictive performance compared with conventional markers.

This study aimed to develop an interpretable machine learning-based scoring system for predicting sepsis and septic shock among febrile patients at emergency department (ED) triage using longitudinal data. This retrospective, single-center study included adult patients, presented to ED of tertiary academic hospital with fever from January 2016 to December 2021. Using the AutoScore framework, we developed a novel scoring system for predicting sepsis and septic shock at the triage stage, incorporating nine variables and a maximum score of 29. The predictive performance of our score was assessed by calculating the area under the receiver operating characteristic curve (AUROC), and its performance was compared with that of two existing scoring systems: the quick Sequential Organ Failure Assessment (qSOFA) and the Modified Early Warning Score (MEWS). Our model incorporated nine variables including initial vital signs, age, baseline platelet count, total bilirubin, and creatinine levels. Among these, initial systolic blood pressure was identified as the most important predictor. AUROC of our model was 0.844 (95

Background: Sepsis remains a leading cause of mortality worldwide. This study evaluated the independent and combined effects of age and chronic comorbidities on clinical outcomes in patients with septic shock. Methods: We conducted a multicenter retrospective observational study to evaluate the factors associated with 28-day mortality in the Korean Shock Society registry between 2015 and 2023. Adults with suspected infection and refractory hypotension or hypoperfusion within 6 h of emergency department (ED) arrival were included. Patients were grouped by age (<50, 50-74, and ≥75 years) and comorbidity status. Comorbidities encompass major chronic conditions including hypertension, diabetes mellitus, malignancy, history of organ transplant, dementia, nursing home residence, chronic disease of cardiac, lung, liver, and kidney. The primary outcome was 28-day mortality. Multivariable logistic regression analysis was used. Results: Among 8787 patients (median age 70.2 years), the 28-day mortality rate was 22.9% (n = 2018). Elderly patients with comorbidities had the highest mortality (27.5%). Additionally, patients aged over 50 with at least one comorbidity accounted for 18% of the total cohort (n = 1605) but accounted for nearly 80% of all 28-day deaths. Although younger patients without comorbidities represented a small subgroup, their mortality was not negligible (7.3%) and was substantially higher with comorbidities (22.2%). Compared with patients <50 years, adjusted odds ratios (aORs) of 28-day mortality were 1.81 (95% CI, 1.08-3.03) for 50-74 years and 3.21 (95% CI, 1.92-5.37) for ≥75. The presence of any comorbidities was independently associated with higher odds of 28-day mortality compared with no comorbidity (aOR 2.67; 95% CI, 1.57-4.54). A significant interaction between age and comorbidity status (p for interaction = 0.008) suggested that the age-related gradient in mortality differed depending on comorbidity burden. Conclusions: Age and comorbidities were both significantly associated with septic shock mortality, and their significant interaction demonstrates effect modification, indicating that the prognostic impact of comorbidities differs by age group and that age-related mortality gradients are influenced by comorbidity burden.

Purpose:This study aimed to identify germline pathogenic/likely pathogenic variants in DNA damage repair genes associated with increased cancer risk in Korean patients with biliary tract cancer and characterize their population-specific patterns. Materials and Methods:In this retrospective multicenter cohort study, we performed germline whole-exome sequencing in 172 Korean patients diagnosed with intrahepatic cholangiocarcinoma (n = 83) or gallbladder cancer (n = 89) between June 2001 and February 2022. Germline variants were analyzed in 210 hereditary cancer genes, and the germline landscape of this cohort was compared with that of global cohorts. Results:Pathogenic/likely pathogenic variants were identified in 24 of 172 (14.0%) patients, predominantly in DNA damage repair genes (18 of 24 [75.0%]). BRCA2 was among the most frequently altered genes, harboring two distinct pathogenic variants (2 of 24 [8.3%]; both cases of intrahepatic cholangiocarcinoma). Of the 24 carriers, five (20.8%) harbored Tier 1-2 variants of potential, tumor-confirmation-dependent therapeutic relevance. Notably, 15 of 24 (62.5%) carriers reported no family cancer history. In population-stratified comparisons across nine biliary tract cancer cohorts (n = 4,018), PMS2 showed a Korean-enriched signal after accounting for heterogeneous gene coverage, whereas TP53 showed only a directional, non-significant increase after multiple-testing correction. Conclusion:The study findings provide reference data for genetic counseling in East Asian patients with biliary tract cancer and suggest that germline testing may warrant consideration regardless of family history.

The development of structured optical character recognition (OCR) systems for Korean medical documents is hindered by strict privacy regulations that limit access to real clinical data. Existing synthetic document frameworks are primarily designed for generic layouts and Western languages, and fail to capture the complex formatting, bilingual terminology, and hierarchical field semantics required for clinical OCR and document-understanding tasks in hospital environments. To develop and validate a privacy-preserving, template-driven pipeline for generating realistic syn- thetic Korean medical documents with structured XML annotations, enabling OCR and document-understanding model training in settings where access to real clinical data is restricted. We designed a seven-stage synthetic document generation pipeline that reconstructs authentic medical document layouts from publicly available references, incorporates clinician-guided template selection, and employs a structured placeholder system encoding 17 medical datatypes. Human-curated templates were populated with exemplar content and tagged using a delimiter-based scheme, after which GPT-4o-mini generated diverse synthetic values for each field. The pipeline produced 1,000 fully synthetic DOCX documents with corresponding XML annotations encoding hierarchical structure, title–answer relationships, and datatype metadata. We evaluated text fidelity, lexical diversity (TTR, MTLD, n-grams, Pareto), and compared the system with existing synthetic document pipelines. A Qwen2.5-VL model was fine-tuned exclusively on the synthetic dataset and evaluated on 50 real-world Korean clinical documents to assess OCR accuracy, information extraction, and table reconstruction. The pipeline achieved 100% placeholder coverage, 99.68% format correctness, and 91.56% semantic accuracy. Lexical diversity analysis demonstrated high variability in free-text fields (MTLD ≈ 456) and expected repetitiveness in structured datatypes. Compared with existing pipelines (SynthText, SynthTiger, DocSynth), the proposed system uniquely supports complex clinical forms, bilingual content, and structured XML annotations. When trained solely on the synthetic data, Qwen2.5-VL achieved the highest real-world performance, with a character accuracy of 86.01%, word accuracy of 78.25%, field accuracy of 67.18%, datatype accuracy of 64.14%, and table reconstruction TEDS score of 71.09%. Models without fine-tuning performed substantially worse. We present the first synthetic data generation pipeline specifically tailored to Korean medical documents, enabling the creation of realistic, richly annotated datasets for structured OCR without exposing patient information. The synthetic data significantly improves OCR performance on real clinical documents, demonstrating the pipeline’s practical utility for hospital deployment and its potential extensibility to other languages and healthcare systems. -

Persons with disabilities experience significant health inequities globally, driven in part by the lack of disability-disaggregated data in national health surveillance systems. In Korea, despite having a well-established administrative infrastructure-most notably the Korean National Disability Registration System-disability-inclusive health data remain limited. The National Health Information Database, which links administrative and health insurance data, offers a rare example of integrated data enabling individual-level analysis by disability status, type, and severity. However, the absence of a legal mandate and limited policy recognition of disability inclusion have hindered the widespread use of disability identifiers across key national health datasets. Korea's experience illustrates both the potential and the limitations of relying solely on administrative definitions, which often fail to reflect the broader spectrum of disability outlined in international frameworks such as the International Classification of Functioning, Disability and Health. To advance equity in health systems, countries must integrate disability into population-based data using standardized, functional definitions. This commentary emphasizes the need for both technical integration and a conceptual shift in how disability is defined, measured, and operationalized in health data systems.

BACKGROUND:Given the increased vulnerability of people with disabilities to poor health outcomes, we evaluated the impact of disability on long-term mortality among tuberculosis (TB) survivors. METHODS:We conducted a nationwide population-based cohort study using the linked national registry databases in the Republic of Korea. The study included 305,055 TB patients diagnosed between 2008 and 2016 who survived at least 1 year. The primary outcome was to compare long-term mortality after TB diagnosis between people with and without disabilities. Long-term mortality was defined as all-cause mortality at least 1 year after TB diagnosis. Cox proportional hazard models were used to evaluate the risk of long-term mortality. Subgroup and sensitivity analyses were performed based on disability type, severity, and cause of death. RESULTS:Disabilities were present in 10.1% of survivors and were associated with higher mortality rates (46.3 vs. 16.3 per 1,000 person-years, P < 0.001). Cox analysis revealed that disabilities increased long-term mortality risk, with severe disabilities posing the highest risk. Respiratory disabilities were strongly linked to deaths both related and unrelated to TB. CONCLUSION:Long-term mortality risk is significantly higher in TB survivors with disabilities.

Polycystic kidney disease (PKD) typically manifests as genetic disease, which is commonly attributed to mutations in PKD genes. In this particular case, however, genetic analysis revealed that the patient’s PKD is linked to a novel, likely pathogenic variant (c.2184del; p.Thr729Leufs*88) in the oral-facial-digital syndrome type I (OFD1) gene. This is the first confirmed genetic diagnosis of mutations in the OFD1 gene in Korea. This investigation emphasizes the critical utility of panel sequencing of PKD in offering precise diagnosis and understanding the genetic profiles of PKD.

Introduction Although genetic testing for hereditary cancers is increasing, data on health attitudes based on genetic pathogenicity are limited. This cohort study aims to establish three subcohorts based on genetic testing results to assess the health impact of genetic variations. This study evaluates changes in participant quality of life (QoL), unmet needs and mental health over time based on their genetic variant status.Methods and analysis This prospective cohort study will recruit 1435 patients with suspected hereditary cancer who have undergone BRCA1/2 or next-generation sequencing (NGS) testing. The study began in July 2023 and will continue until December 2027. By 2026, participants will be surveyed up to four times annually during their outpatient visits. The survey consists of 342 items across 5 domains: comorbidities (96), health behaviours (80), QoL (41), unmet needs (75) and mental health (50). Data were collected using 11 validated surveys. In addition, information on the chronic diseases, cancer diagnoses, medical history and treatment history of participants will be extracted from their electronic medical records to analyse their health status and cancer treatment experiences. Genetic variant data from BRCA1/2 and NGS will be used to classify participants into three subcohorts: pathogenic variants, variants of uncertain significance and undetectable mutations. A three-generation pedigree that includes details such as the year of cancer diagnosis, age at diagnosis, cancer type, survival status of family members and age at death will be constructed for each participant. The collected data will be linked to secondary sources such as cancer registries and National Health Insurance Service data to provide a comprehensive analysis of the impact of hereditary cancer on health and survival.Ethics and dissemination The study protocol was approved by all the Ethics Committees: the National Cancer Center IRB (NCC2023-0179), the Samsung Medical Center IRB (SMC2023-09-057), the Yonsei University Health System, Severance Hospital IRB (4-2023-0627), the Hanyang University Guri Hospital IRB (GURI2023-08-021) and the Keimyung University IRB (DSMC IRB 2024-05-048). The study outcomes will be disseminated through conference presentations, peer-reviewed publications and social media.Trial registration number KCT0009460.

Checkpoint kinase 2 (CHEK2) encodes a serine/threonine kinase involved in the DNA damage response through ATM-Chk2-p53 signaling. Its function in maintaining genomic stability classifies it as a tumor suppressor. Heterozygous germline pathogenic variants in CHEK2 are associated with a moderate increase in lifetime risk of breast and prostate cancer. This study assessed the prevalence of CHEK2 variants globally, with a focus on East Asian and Korean populations, for which data have remained limited. We analyzed 125,748 exomes from the Genome Aggregation Database (gnomAD), including 9,197 East Asians, along with additional data from 5,305 individuals in the Korean Variant Archive, 3,617 in Korea4K, and 1,722 in the Korean Reference Genome Database. All CHEK2 variants were classified according to guidelines established by the American College of Medical Genetics, Genomics, and Clinical Genome Resources. The global prevalence of CHEK2 variants was 0.76%, with the highest observed in the Finnish population (2.04%) and the lowest in East Asians (0.11%). By integrating data from Korean genomic databases and gnomAD, representing a total of 12,553 Korean individuals, the overall prevalence in the Korean population was estimated at 0.13%. These findings represent the first integrated estimate of CHEK2 variant frequency in Koreans using multiple population-specific genomic datasets. The results provide a useful reference for future studies and highlight the need for region-specific genetic research to inform counseling and hereditary cancer risk management.

This study aimed to develop and validate a transformer-based early warning score (TEWS) system for predicting adverse events (AEs) in the emergency department (ED). We conducted a retrospective study analyzing adult ED visits at a tertiary hospital. The TEWS was developed to predict five AEs within 24 h: vasopressor use, respiratory support, intensive care unit admission, septic shock, and cardiac arrest. Performance was evaluated and compared using the area under the receiver operating characteristic curve (AUROC) and bootstrap-based t-test. External validation was performed using the Marketplace for Medical Information in Intensive Care (MIMIC)-IV-ED database. Transfer learning was applied using 1% and 5% of the external data. A total of 414,748 patients was analyzed in the development cohort (AEs, 3.7%), and 410,880 patients (AEs, 6.7%) were included in the external validation cohort. Compared to the modified early warning score (MEWS), the TEWS incorporating 13 variables and the vital signs-only TEWS demonstrated superior prognostic performance across all AEs. The AUROC ranged from 0.833 to 0.936 for TEWS and 0.688 to 0.874 for MEWS. In external validation, the TEWS also showed acceptable discrimination with AUROC values of 0.759 to 0.905. Transfer learning significantly improved the performance, increasing AUROC values to 0.846–0.911. The TEWS system was successfully integrated into the electronic health record (EHR) system of the study hospital, providing real-time risk assessment for ED patients. We developed and validated an artificial intelligence-based early warning score system that predicts multiple adverse outcomes in the ED and was successfully integrated into the EHR system.

TSC1 and TSC2 encode hamartin and tuberin, key regulators of the mTOR pathway that controls cell growth and proliferation. Germline pathogenic variants in these genes cause tuberous sclerosis complex (TSC), a multisystem disorder frequently associated with neurological tumors such as subependymal giant cell astrocytomas (SEGAs) and cortical tubers. However, population-level data on TSC1 and TSC2 variant frequencies remain scarce, particularly in East Asian populations. We analyzed exome sequencing data from 125,748 individuals in the Genome Aggregation Database (gnomAD), including 9,197 East Asians, as well as Korean-specific datasets comprising 5,305 individuals from the Korean Variant Archive (KOVA), 3,617 from Korea4K, and 1,722 from the Korean Reference Genome Database (KRGDB). Variants in TSC1 and TSC2 were annotated and classified according to ACMG/AMP and ClinGen guidelines, focusing on pathogenic and likely pathogenic variants. The global carrier frequency was estimated at 0.005% for TSC1 and 0.007% for TSC2. Among populations, the TSC1 carrier frequency was highest in Ashkenazi Jewish (0.040%) followed by East Asians (0.011%). TSC2 carrier frequency was also elevated in East Asians (0.011%), second only to Latinos (0.012%). In the integrated Korean dataset (n = 12,553), the estimated carrier frequencies were 0.032% for TSC1 and 0.032% for TSC2, indicating a notably higher prevalence in Koreans compared to other East Asian subgroups. This study provides the first comprehensive, population-based estimate of TSC1 and TSC2 pathogenic variant frequencies in Koreans. These findings highlight a potentially elevated burden of hereditary TSC in this population and underscore the need for population-specific genomic data to support accurate risk assessment, early detection, and clinical management of neuro-oncologic manifestations associated with TSC.

BACKGROUND/OBJECTIVES:There has been a notable lack of effort to evaluate the nutritional status of persons with disabilities objectively and identify their nutritional challenges. This study aimed to assess whether nutritional inadequacies and imbalances are more pronounced in persons with disabilities compared to those without, using data from the 6th Korea National Health and Nutrition Examination Survey (2013). SUBJECTS/METHODS:The participants were classified into 387 persons with disabilities and 4,909 without disabilities. The 15 types of disabilities were categorized into 5 groups, and disability severity was classified as severe or mild. Nutrient intake and nutritional status were assessed using 24-h dietary recall data. RESULTS:Compared to persons without disabilities, those with disabilities had significantly lower nutrient adequacy ratios for 8 of 10 nutrients, the exceptions being vitamins B1 and C (all P < 0.05). The likelihood of low overall nutrient adequacy based on the mean adequacy ratio was significantly higher among persons with disabilities (odds ratio [OR], 1.50; 95% confidence interval [CI], 1.11-2.04), with particularly elevated odds for those with severe, visual, and internal organ disabilities. Persons with severe disabilities had higher odds of consuming a diet with low nutrient density (OR, 1.59; 95% CI, 1.05-2.40) and of experiencing nutritional deficiency (OR, 2.21; 95% CI, 1.02-4.75) than those without disabilities. Additionally, the odds of consuming a high-carbohydrate diet, with > 70% of daily energy intake from carbohydrates, was 3.63-fold higher (95% CI, 1.40-9.42) among persons with mental disabilities. CONCLUSION:Persons with disabilities faced significant nutritional inadequacies and imbalances compared to those without disabilities. Targeted nutrition interventions and disability-inclusive nutrition policies and practices are urgently needed to reduce these disparities.

OBJECTIVE:To develop a scale to predict refractory septic shock (SS) based on clinical variables recorded during initial evaluations of patients. METHODS:Multicenter retrospective study of data for patients with suspected infection registered in the Marketplace for Medical Information in Intensive Care (MIMIC-IV). These data were used for the development and internal validation of the refractory SS scale (RSSS). For external validation, we used retrospective data for 2 cohorts: 1) patients diagnosed with SS in an emergency department (ED cohort) whose data were registered in a Korean SS registry, and 2) patients diagnosed with SS in 6 hospital intensive care units (ICU cohort). A machine-learning automatic clinical scoring system (AutoScore) was used in the development phase. The performance of the RSSS in the validation cohorts was assessed with the area under the receiver operating characteristic curve (AUROC) for each. The primary outcome was the development of refractory SS within 24 hours of ICU admission. Refractory SS was defined by the need for a norepinephrine-equivalent dose greater than 0.5 µg/kg/min. RESULTS:We collected data for 29 618 patients from the MIMIC-IV registry, 3113 patients for the ED cohort, and 1015 for the ICU cohort. The RSSS had 6 predictors: serum lactate level, systolic blood pressure, heart rate, temperature, arterial pH, and leukocyte count. The scale's AUROCs were as follows: 0.873 (95% CI, 0.846-0.900) in the internal validation, 0.705 (95% CI, 0.678-0.733) in the ED cohort on arrival, 0.781 (95% CI, 0.757-0.805) in the ED cohort at the moment of diagnosing hypoperfusion or hypotension, and 0.822 (95% CI, 0.787-0.857) in the ICU cohort. Calibration was acceptable in all the cohorts. CONCLUSIONS:The RSSS had adequate diagnostic accuracy in multiple cohorts of patients diagnosed in the ED and ICU.

Background: Accurate documentation of cardiopulmonary resuscitation (CPR) is essential. However, traditional methods, particularly handwriting, often introduce errors and increase the workload of the medical staff. This study aimed to describe the process of developing a tablet application for documenting CPR and to evaluate its accuracy in comparison with a paper-based method. Methods: We organized a multidisciplinary team of medical professionals, developers, and designers. We used a participative human-centered design (HCD) approach that consisted of discovering, defining, developing and delivering solutions. We conducted a simulation study to compare the accuracy of the CPR documentation application with that of the handwriting method, focusing on documentation completeness and temporal fidelity. We evaluated the usability of the application using a System Usability Scale (SUS) and semi-structured interviews. Results: We developed the "CPReCoder" in accordance with the HCD process. The study application consists of two screens: a CPR recording screen and a reporting screen. The CPR recording screen is divided into three zones: zone 1 (patient and prehospital area), zone 2 (CPR code button area) and zone 3 (time information and log area). In the simulation study, the documentation completeness of the "CPReCoder" was significantly higher than that of the handwritten record (96.8% vs. 88.1%, p < 0.001). Both approaches exhibit comparable temporal fidelity. The SUS score of the application was 87.9 points, indicating excellent usability. According to the responses in the interviews, the main benefit of CPReCoder was its ability to reduce workload. Conclusions: We described the process of creating a CPR recording application. Use of the application resulted in a more complete documentation than the handwriting method, and its usability was excellent.