Teacher-Explorer-Student Learning: A Novel Learning Method for Open Set Recognition.

When an unknown example, one that was not seen during training, appears, most recognition systems usually produce overgeneralized results and determine that the example belongs to one of the known classes. To address this problem, teacher-explorer-student (T/E/S) learning, which adopts the concept of open set recognition (OSR) to reject unknown samples while minimizing the loss of classification performance on known samples, is proposed in this study. In this novel learning method, the overgeneralization of deep-learning classifiers is significantly reduced by exploring various possibilities for unknowns. The teacher network extracts hints about unknowns by distilling the pretrained knowledge about knowns and delivers this distilled knowledge to the student network. After learning the distilled knowledge, the student network shares its learned information with the explorer network. Next, the explorer network shares its exploration results by generating unknown-like samples and feeding those samples to the student network. As this alternating learning process is repeated, the student network experiences a variety of synthetic unknowns, reducing overgeneralization. The results of extensive experiments show that each component proposed in this article significantly contributes to improving OSR performance. It is found that the proposed T/E/S learning method outperforms current state-of-the-art methods.

Corrigendum: Combination effect of engineered endolysin EC340 with antibiotics.

[This corrects the article DOI: 10.3389/fmicb.2022.821936.].

Tumor mutation burden in gastro-entero-pancreatic-neuroendocrine neoplasms.

Background: As rare tumors, there are limited treatment options for neuroendocrine neoplasms (NENs). Recently, microsatellite instability (MSI) and tumor mutation burden (TMB) have been emerging as potential biomarkers in various tumors. However, there is a lack of research on the use of these biomarkers in gastro-entero-pancreatic (GEP)-NENs. Methods: We analyzed 31 patients diagnosed with GEP-NEN between 2013 to 2022. The TMB and MSI analyses using next-generation sequencing (NGS) were performed for all patients. The TruSightTM Oncology 500 assay from Illumina was used as the NGS panel. Results: Out of the 31 patients analyzed, the most frequent primary origin was the pancreas (12 patients, 38.7%), followed by the stomach (4 patients, 12.9%), gallbladder (4 patients, 12.9%), rectum (7 patients, 22.6%), small bowel (2 patients, 6.5%), and bile duct (1 patient, 3.2%). Among these patients, 19 (61.3%) were diagnosed with well-differentiated neuroendocrine tumors, with grade 2 being the most common (15 patients, 48.4%), followed by grade 3 (3 patients, 9.7%) and grade 1 (1 patient, 3.2%). Neuroendocrine carcinoma was confirmed in 12 patients (38.7%). The median number of metastases was 2.0 [interquartile range (IQR), 1.0-3.0], and the liver was the most common site of metastasis (23 patients, 74.2%). The median TMB was 4.7 (IQR, 3.1-6.3) mutations/Mb, and all tumors were classified as microsatellite stability (MSS). Only one patient had a high TMB (266.4 mutations/Mb), which was a grade 3 neuroendocrine tumor originating from the pancreas. The TMB value did not vary depending on the primary tumor site or World Health Organization (WHO) grade. Conclusions: This analysis showed that, despite very low incidence, there are GEP-NENs with high TMB. For precision medicine, testing for MSI and TMB is needed for this tumor type.

Collective Decision of One-vs-Rest Networks for Open-Set Recognition.

Unknown examples that are unseen during training often appear in real-world pattern recognition tasks, and an intelligent self-learning system should be able to distinguish between known examples and unknown examples. Accordingly, open-set recognition (OSR), which addresses the problem of classifying knowns and identifying unknowns, has recently been highlighted. However, conventional deep neural networks (DNNs) using a softmax layer are vulnerable to overgeneralization, producing high confidence scores for unknowns. In this article, we propose a simple OSR method that is based on the intuition that the OSR performance can be maximized by setting strict and sophisticated decision boundaries that reject unknowns while maintaining satisfactory classification performance for knowns. For this purpose, a novel network structure, in which multiple one-vs-rest networks (OVRNs) follow a convolutional neural network (CNN) feature extractor, is proposed. Here, an OVRN is a simple feedforward neural network that is designed to assign confidence scores that are lower than those in the softmax layer to unknown samples so that unknown samples can be more effectively separated from known classes. Furthermore, the collective decision score is modeled by combining the multiple decisions reached by the OVRNs to alleviate overgeneralization. Extensive experiments were conducted on various datasets, and the experimental results show that the proposed method performs significantly better than the state-of-the-art methods by effectively reducing overgeneralization. The code is available at https://github.com/JaeyeonJang/Openset-collective-decision.

Eradication of drug-resistant Acinetobacter baumannii by cell-penetrating peptide fused endolysin.

Antimicrobial agents targeting peptidoglycan have shown successful results in eliminating bacteria with high selective toxicity. Bacteriophage encoded endolysin as an alternative antibiotics is a peptidoglycan degrading enzyme with a low rate of resistance. Here, the engineered endolysin was developed to defeat multiple drug-resistant (MDR) Acinetobacter baumannii. First, putative endolysin PA90 was predicted by genome analysis of isolated Pseudomonas phage PBPA. The His-tagged PA90 was purified from BL21(DE3) pLysS and tested for the enzymatic activity using Gram-negative pathogens known for having a high antibiotic resistance rate including A. baumannii. Since the measured activity of PA90 was low, probably due to the outer membrane, cell-penetrating peptide (CPP) DS4.3 was introduced at the N-terminus of PA90 to aid access to its substrate. This engineered endolysin, DS-PA90, completely killed A. baumannii at 0.25 µM, at which concentration PA90 could only eliminate less than one log in CFU/ml. Additionally, DS-PA90 has tolerance to NaCl, where the ∼50% of activity could be maintained in the presence of 150 mM NaCl, and stable activity was also observed with changes in pH or temperature. Even MDR A. baumannii strains were highly susceptible to DS-PA90 treatment: five out of nine strains were entirely killed and four strains were reduced by 3-4 log in CFU/ml. Consequently, DS-PA90 could protect waxworm from A. baumannii-induced death by ∼70% for ATCC 17978 or ∼44% for MDR strain 1656-2 infection. Collectively, our data suggest that CPP-fused endolysin can be an effective antibacterial agent against Gram-negative pathogens regardless of antibiotics resistance mechanisms.

Combination Effect of Engineered Endolysin EC340 With Antibiotics.

Bacteriophage lysins, also known as endolysins or murein hydrolases, are hydrolytic enzymes produced by bacteriophages during the final stage of the lytic cycle to enable cleavage through the host's cell wall, thus allowing the phages to burst out of their host bacteria after multiplication inside them. When applied externally to Gram-negative bacteria as recombinant proteins, lysins cannot easily reach the cell wall due to the presence of an outer membrane (OM). In this study, endolysin EC340 obtained from phage PBEC131 infecting Escherichia coli was engineered for improved OM permeability and increased activity against Gram-negative bacteria. The engineered endolysin, LNT113, was tested for potential synergistic effects with standard-of-care antibiotics. A synergistic effect was demonstrated with colistin, while an additive effect was seen with meropenem, tigecycline, chloramphenicol, azithromycin, and ciprofloxacin. Neither ceftazidime nor kanamycin showed any synergy or additive effects with the LNT113 endolysin. Moreover, synergy and additive effects could not be generalized by antibiotic class, OM traverse mechanism, molecular weight, or the bactericidal nature of each antibiotic tested.

Survival impact of adherence to tyrosine kinase inhibitor in chronic myeloid leukemia.

BACKGROUND/AIMS: Adherence to tyrosine kinase inhibitors (TKIs) has become a critical aspect of care in chronic myeloid leukemia (CML). We aimed to examine the association of TKI adherence with overall survival (OS) outcomes in Korean patients diagnosed with CML and treated with TKIs using data from the National Health Information Database. METHODS: This study included 2,870 CML patients diagnosed between 2005 and 2013. Drug adherence was evaluated according to the medication possession ratio (MPR) and classified as high adherence (i.e., MPR ≥ 0.95 [upper 50%]), moderate adherence (i.e., MPR ≥ 0.68 and < 0.95 [middle 25%]), and low adherence (i.e., MPR < 0.68 [lower 25%]). RESULTS: The median MPR was 0.95 (range, 0 to 4.67). Male sex (p = 0.003), age < 70 years (p < 0.001), high income (≥ 30%, p < 0.001), and maintaining frontline TKI (< 0.001) were associated with better adherence. Adherence to dasatinib was the lowest (vs. imatinib or nilotinib, p < 0.001). Compared with high MPR patients, those with moderate MPR (hazard ratio [HR], 4.90; 95% confidence interval [CI], 3.87 to 6.19; p < 0.001) and low MPR (HR, 11.6; 95% CI, 9.35 to 14.42; p < 0.001) had poorer OS. CONCLUSION: Adherence to TKI treatment is an important factor predicting survival outcomes in Korean CML patients. Male sex, age < 70 years, high income, and maintaining frontline TKI are associated with high adherence to TKI. Thus, those without these characteristics should be closely monitored for treatment adherence.

TimesVector-Web: A Web Service for Analysing Time Course Transcriptome Data with Multiple Conditions.

From time course gene expression data, we may identify genes that modulate in a certain pattern across time. Such patterns are advantageous to investigate the transcriptomic response to a certain condition. Especially, it is of interest to compare two or more conditions to detect gene expression patterns that significantly differ between them. Time course analysis can become difficult using traditional differentially expressed gene (DEG) analysis methods since they are based on pair-wise sample comparison instead of a series of time points. Most importantly, the related tools are mostly available as local Software, requiring technical expertise. Here, we present TimesVector-web, which is an easy to use web service for analysing time course gene expression data with multiple conditions. The web-service was developed to (1) alleviate the burden for analyzing multi-class time course data and (2) provide downstream analysis on the results for biological interpretation including TF, miRNA target, gene ontology and pathway analysis. TimesVector-web was validated using three case studies that use both microarray and RNA-seq time course data and showed that the results captured important biological findings from the original studies.

Dependency on Smartphone Use and Its Association with Anxiety in Korea.

OBJECTIVE: South Korea has the highest rate of smartphone ownership worldwide, which is a potential concern given that smartphone dependency may have deleterious effects on health. We investigated the relationship between smartphone dependency and anxiety. METHODS: Participants included 1,236 smartphone-using students (725 men and 511 women) from six universities in Suwon, South Korea. Participants completed measures of smartphone use, smartphone dependency, anxiety, and general characteristics (i.e., demographic, health-related, and socioeconomic characteristics). To measure smartphone dependency and anxiety, we used questionnaires of Yang's test developed from Young's Internet Addiction Test and Zung's Self-Rating Anxiety Scale. We used multiple logistic regression to determine the association between smartphone dependency and anxiety after adjusting for relevant factors. RESULTS: On a scale from 25 to 100, with higher scores on the smartphone dependency test indicating greater dependency, women were significantly more dependent on smartphones than were men (mean smartphone dependency score: 50.7 vs. 56.0 for men and women, respectively, p<0.001). However, the amount of time spent using smartphones and the purpose of smartphone use affected smartphone dependency in both men and women. Particularly, when daily use time increased, smartphone dependency showed an increasing trend. Compared with times of use <2 hours vs. ≥6 hours, men scored 46.2 and 56.0 on the smartphone dependency test, while women scored 48.0 and 60.4, respectively (p<0.001). Finally, for both men and women, increases in smartphone dependency were associated with increased anxiety scores. With each one-point increase in smartphone dependency score, the risk of abnormal anxiety in men and women increased by 10.1% and 9.2%, respectively (p<0.001). CONCLUSION: Among this group of university students in South Korea, smartphone dependency appeared to be associated with increased anxiety. Standards for smartphone use might help prevent deleterious health effects.

Incidences of Waterborne and Foodborne Diseases After Meteorologic Disasters in South Korea.

BACKGROUND: Climate change could increase the number of regions affected by meteorologic disasters. Meteorologic disasters can increase the risk of infectious disease outbreaks, including waterborne and foodborne diseases. Although many outbreaks of waterborne diseases after single disasters have been analyzed, there have not been sufficient studies reporting comprehensive analyses of cases occurring during long-term surveillance after multiple disasters, which could provide evidence of whether meteorologic disasters cause infectious disease outbreaks. OBJECTIVES: This study aimed to assess the nationwide short-term changes in waterborne and foodborne disease incidences after a meteorologic disaster. METHODS: We analyzed cases after all 65 floods and typhoons between 2001 and 2009 using the Korean National Emergency Management Agency's reports. Based on these data, we compared the weekly incidences of Vibrio vulnificus septicemia (VVS), shigellosis, typhoid fever, and paratyphoid fever before, during, and after the disasters, using multivariate Poisson regression models. We also analyzed the interactions between disaster characteristics and the relative risk of each disease. FINDINGS: Compared with predisaster incidences, the incidences of VVS and shigellosis were 2.49-fold (95% confidence interval, 1.47-4.22) and 3.10-fold (95% confidence interval, 1.21-7.92) higher, respectively, the second week after the disaster. The incidences of VVS and shigellosis peaked the second week postdisaster and subsequently decreased. The risks of typhoid and paratyphoid fever did not significantly increase throughout the 4 weeks postdisaster. The daily average precipitation interacted with VVS and shigellosis incidences, whereas disaster type only interacted with VVS incidence patterns. CONCLUSIONS: The incidences of VVS and shigellosis were associated with meteorologic disasters, and disaster characteristics were associated with the disease incidence patterns postdisaster. These findings provide important comprehensive evidence to develop and support policies for managing and protecting public health after meteorologic disasters.

Psychological Distress in Young Adult Males with Atopic Dermatitis: A Cross-Sectional Study.

The relationship between atopic dermatitis (AD) and psychological distress has been well established for children and adolescents. However, it is unclear whether this relationship exists in young adults. This study aimed to assess the relationship between AD and psychological distress in young male adults in South Korea. A cross-sectional study was conducted using regional conscription data from 2008 to 2012. A dermatologist diagnosed AD based on historical and clinical features, and determined severity using the eczema area and severity index. A psychiatrist used medical records, an interview, and a psychological test to examine psychological distress (depression, anxiety, and somatization). The relationship between psychological distress and AD was assessed by multivariate logistic regression analyses. Among the 120,508 conscripts, 1517 (1.2%) presented with AD. The odds of having each type of psychological distress were significantly greater for individuals with AD compared with those without AD. The adjusted odds ratios for depression, anxiety, and somatization were 1.79 (95% CI 1.40-2.29), 1.38 (95% CI 1.08-1.76), and 1.75 (95% CI 1.40-2.20), respectively. Moderate-to-severe AD was significantly related to depression and somatization to a greater extent compared with mild AD. Depression, anxiety, and somatization are strongly and independently associated with AD in young adult males. Early treatment of skin inflammation might modify the risk of psychiatric problems. Prospective cohort studies are needed to verify causal relationships.

The effect of pre-existing medical conditions on heat stroke during hot weather in South Korea.

BACKGROUND: Heat stroke contributes considerably to morbidity and mortality in hot weather, but it is unclear whether pre-existing medical conditions increase the risk of heat stroke. The purpose of this study was to assess the association between heat stroke and pre-existing medical conditions in South Korea. METHODS: A case-control study was conducted using data from a nationwide surveillance system for heat illnesses in 2012. Individuals with heat stroke were identified and compared to control subjects with mild heat illness such as heat exhaustion, heat edema, heat cramps, and heat syncope. Categories of pre-existing medical conditions included cardio/cerebrovascular disease, respiratory disease, neuropsychiatric disorder, and diabetes mellitus. Associations between heat stroke and pre-existing medical conditions were assessed by multivariate logistic regression analyses. RESULTS: Of 968 eligible patients, 178 (18.4%) presented with heat stroke. Patients with pre-existing medical conditions were relatively greater in the heat stroke group than in the control group (40.4% and 23.9%, respectively). Neuropsychiatric disorder was associated with an increased odds of heat stroke after adjustment for covariates (adjusted odds ratio, 7.69; 95% CI, 4.06-14.54). There were no significant relationships between heat stroke and other medical conditions (cardio/cerebrovascular disease: 0.66, 0.40-1.06; respiratory disease: 1.44, 0.35-5.89; diabetes mellitus: 1.16, 0.58-2.34). CONCLUSIONS: Pre-existing neuropsychiatric disorder was associated with an increased risk of heat stroke. National strategies and clinical guidance for such patients should be initiated to prevent fatal events.

Overview of the development of the Korean exposure factors handbook.

A set of exposure factors that reflects the characteristics of individual behavior capable of influencing exposure is essential for risk and exposure assessment. In 2007, the Korean Exposure Factors Handbook was, therefore, issued, driven by the need to develop reliable exposure factors representing the Korean population. The purpose of this study was to overview the development process of the Korean Exposure Factors Handbook and major recommended exposure values for the Korean population to allow information exchanges and comparison of recommended values among nations. The researchers reviewed the domestic data that could be used in the development of exposure factors, confirmed a knowledge gap, and set a priority of development by phases. A methodology to measure exposure factors was established to develop measuring techniques and test their validity. Data were processed or a survey was conducted according to the availability of data. The study thus produced recommended values for 24 exposure factors grouped by general exposure factors, food ingestion factors, and activity factors by setting up a database of exposure factors and carrying out statistical analysis. The study has significantly contributed to reducing the potential uncertainty of the risk and exposure assessment derived by the application of foreign data or research findings lacking representativeness or grounds by developing a set of exposure factors reflecting the characteristics of the Korean people. It will be necessary to conduct revisions in light of the changing statistical values of national data and the exposure factors based on Korean characteristics.

Food ingestion factors of the Korean exposure factors handbook.

The purpose of this study was to establish food ingestion factors needed to assess exposure to contaminants through food ingestion. The study reclassified the raw data of the Korean National Health and Nutrition Examination Survey in 2001 into 12 subcategories including grain products, meat products, fish and shellfish, and vegetables for international comparability of exposure evaluation. The criteria for food intake calculation were unified according to the characteristics of food groups, and recommended values for food ingestion factors were calculated through moisture correction and recategorization of cooked, processed, and mixed foods for each group. The average intake rate for grain and grain products was 6.25 g/kg-d per capita and the men's intake rate was approximately 8% higher than that of the women. The average intake rate of meat and meat products was 1.62 g/kg-d per capita and the men's intake rate was 30% higher than that of the women, on average. The average intake rate of fish and shellfish was 1.53 g/kg-d per capita, and the age groups of 1 to 2 and 3 to 6 recorded higher capita intake rates than other age groups, 2.62 g/kg-d and 2.25 g/kg-d, respectively. The average intake rate of vegetables was 6.47 g/kg-d per capita, with the age group of 1 to 2 recording the highest per capita intake rate of 9.79 g/kg-d and that of 13 to 19 recording the lowest mean. The study also offers recommended values for food ingestion factors of other food groups by gender, age, and region. The food ingestion exposure factors will need future updates in consideration of ongoing changes in food consumption behavior.

General factors of the Korean exposure factors handbook.

Risk assessment considers the situations and characteristics of the exposure environment and host. Various physiological variables of the human body reflects the characteristics of the population that can directly influence risk exposure. Therefore, identification of exposure factors based on the Korean population is required for appropriate risk assessment. It is expected that a handbook about general exposure factors will be used by professionals in many fields as well as the risk assessors of the health department. The process of developing the exposure factors handbook for the Korean population will be introduced in this article, with a specific focus on the general exposure factors including life expectancy, body weight, surface area, inhalation rates, amount of water intake, and soil ingestion targeting the Korean population. The researchers used national databases including the Life Table and the 2005 Time Use Survey from the National Statistical Office. The anthropometric study of size in Korea used the resources provided by the Korean Agency for Technology and Standards. In addition, direct measurement and questionnaire surveys of representative samples were performed to calculate the inhalation rate, drinking water intake, and soil ingestion.

Activity factors of the Korean exposure factors handbook.

Exposure factors based on the Korean population are required for making appropriate risk assessment. It is expected that handbooks for exposure factors will be applied in many fields, as well as by health department risk assessors. The present article describes the development of an exposure factors handbook that specifically focuses on human activities in situations involving the possible risk of exposure to environmental contaminants. We define majour exposure factors that represent behavioral patterns for risk assessment, including time spent on routine activities, in different places, on using transportation, and engaged in activities related to water contact including swimming, bathing and washing. Duration of residence and employment are also defined. National survey data were used to identify recommended levels of exposure factors in terms of time spent on routine activities and period of residence and employment. An online survey was conducted with 2073 subjects who were selected using a stratified random sampling method in order to develop a list of exposure factors for the time spent in different places and in performing water-related activities. We provide the statistical distribution of the variables, and report reference levels of average exposure based on the reliable data in our exposure factors handbook.

The Long and Short Incubation Periods of Plasmodium vivax Malaria in Korea: The Characteristics and Relating Factors.

BACKGROUND: The cases of Plasmodium vivax malaria in Korea are mixed with long and short incubation periods. This study aims to define clinico-epidemiologic chracteristcs of Plasmodium vivax malaria in Korea. MATERIALS AND METHODS: We selected the civilian cases infected with P. vivax malaria in Korea from the epidemiological investigation data of 2001 to 2010, whose incubation periods could be estimated. The long and short incubation periods were defined by duration of infection and onset time, and the cases were compared by demographic factors and clinical symptom, infection and onset time. The correlation was analyzed between the proportion of cases in the infected region with the long incubation period and meteorological factors along with latitude. RESULTS: The length of the mean short and long incubation periods for the cases were 25.5 days and 329.4 days, respectively. The total number of the study subjects was 897, and the number cases of short and long incubation periods was 575 (64.1%) and 322 (35.9%), respectively. The aspect of incubation period showed a significant difference by region of infection; there was a higher proportion of long incubation period infection cases in Gangwon-do than in Gyeonggi-do and Incheon. The proportion of long incubation period cases showed significant correlation with latitude and temperature of August and September of the infected regions. CONCLUSIONS: Incubation period of P. vivax malaria in Korea showed significant difference by infected region, infection and onset time and the proportion of long incubation period cases showed significant correlation with latitude and meteorological factors of the infected regions.

Socio-demographic characteristics and leading causes of death among the casualties of meteorological events compared with all-cause deaths in Korea, 2000-2011.

OBJECTIVES: This study investigated the socio-demographic characteristics and medical causes of death among meteorological disaster casualties and compared them with deaths from all causes. METHODS: Based on the death data provided by the National Statistical Office from 2000 to 2011, the authors analyzed the gender, age, and region of 709 casualties whose external causes were recorded as natural events (X330-X389). Exact matching was applied to compare between deaths from meteorological disasters and all deaths. RESULTS: The total number of deaths for last 12 years was 2 728 505. After exact matching, 642 casualties of meteorological disasters were matched to 6815 all-cause deaths, which were defined as general deaths. The mean age of the meteorological disaster casualties was 51.56, which was lower than that of the general deaths by 17.02 (p<0.001). As for the gender ratio, 62.34% of the meteorological event casualties were male. While 54.09% of the matched all-cause deaths occurred at a medical institution, only 7.6% of casualties from meteorological events did. As for occupation, the rate of those working in agriculture, forestry, and fishery jobs was twice as high in the casualties from meteorological disasters as that in the general deaths (p<0.001). Meteorological disaster-related injuries like drowning were more prevalent in the casualties of meteorological events (57.48%). The rate of amputation and crushing injury in deaths from meteorological disasters was three times as high as in the general deaths. CONCLUSIONS: The new information gained on the particular characteristics contributing to casualties from meteorological events will be useful for developing prevention policies.

Assessment of exposure to heavy metals and health risks among residents near abandoned metal mines in Goseong, Korea.

Metal contamination from mining activity is of great concern because of potential health risks to the local inhabitants. In the present study, we investigated the levels of Cd, Cu, As, Pb, and Zn in environmental samples and foodstuffs grown in the vicinity of the mines in Goseong, Korea, and evaluated potential health risks among local residents. Soils near the mines exceeded the soil quality standard values of Cu, As, and Zn contamination. The concentrations of Cd, Cu, Pb, and Zn in crop samples collected from the study area were significantly higher than those of the reference area. Some rice samples collected from the study area exceeded the maximum permissible level of 0.2 mg Cd/kg. The intake of rice was identified as a major contributor (≥75%) to the estimated daily intake among the residents. The average estimated daily intakes of metals were, however, below the provisional tolerable daily intake.

The effects of temperature on heat-related illness according to the characteristics of patients during the summer of 2012 in the Republic of Korea.

OBJECTIVES: This study was conducted to investigate the relationship between heat-related illnesses developed in the summer of 2012 and temperature. METHODS: The study analyzed data generated by a heat wave surveillance system operated by the Korea Centers for Disease Control and Prevention during the summer of 2012. The daily maximum temperature, average temperature, and maximum heat index were compared to identify the most suitable index for this study. A piecewise linear model was used to identify the threshold temperature and the relative risk (RR) above the threshold temperature according to patient characteristics and region. RESULTS: The total number of patients during the 3 months was 975. Of the three temperature indicators, the daily maximum temperature showed the best goodness of fit with the model. The RR of the total patient incidence was 1.691 (1.641 to 1.743) per 1℃ after 31.2℃. The RR above the threshold temperature of women (1.822, 1.716 to 1.934) was greater than that of men (1.643, 1.587 to 1.701). The threshold temperature was the lowest in the age group of 20 to 64 (30.4℃), and the RR was the highest in the ≥65 age group (1.863, 1.755 to 1.978). The threshold temperature of the provinces (30.5℃) was lower than that of the metropolitan cities (32.2℃). Metropolitan cities at higher latitudes had a greater RR than other cities at lower latitudes. CONCLUSIONS: The influences of temperature on heat-related illnesses vary according to gender, age, and region. A surveillance system and public health program should reflect these factors in their implementation.