• Journal of the Korean Data and Information Science Society
• /
• v.19 no.3
• /
• pp.735-749
• /
• 2008

This article suggests application of logit model for analysis of multicategory responses. Referring to the reference category, characteristic of each category is obtained from analysis of polytomous logit model. With National Pension data it is illustrated that application of logit model helps it possible to find significant factors which may not be found only with polytomous logit model. Application of the logit model is done by reducing the number of categories. Categories are grouped into the former and the latter group according to reference category. Extra finding of significant factor was possible from logistic regression analysis for the two groups after removing the reference category. It is expected that this application would be helpful for finding information and characteristics on ordered multicategory responses where the proportional odds model does not fit.

• Communications for Statistical Applications and Methods
• /
• v.16 no.4
• /
• pp.697-705
• /
• 2009

The Pearson chi-squared statistic or the deviance statistic is widely used in assessing the goodness-of-fit of the generalized linear models. But these statistics are not proper in the situation of continuous explanatory variables which results in the sparseness of cell frequencies. We propose a goodness-of-fit test statistic for the cumulative logit models with ordinal responses. We consider the grouping of a dataset based on the ordinal scores obtained by fitting the assumed model. We propose the Pearson chi-squared type test statistic, which is obtained from the cross-classified table formed by the subgroups of ordinal scores and the response categories. Because the limiting distribution of the chi-squared type statistic is intractable we suggest the parametric bootstrap testing procedure to approximate the distribution of the proposed test statistic.

• Journal of the Korean Data and Information Science Society
• /
• v.15 no.2
• /
• pp.367-378
• /
• 2004

Recently logistic regression is popular in a variety of fields so that a number of statistical packages are developed for analyzing the logistic regression. This paper briefly considers the several types of logistic regression models used depending on different types of data. In addition, when four statistical packages (MINTAB, SAS, SPSS and STATA) are used to apply logistic regression models to the real fields respectively, their scope and characteristics are investigated.

• Communications for Statistical Applications and Methods
• /
• v.29 no.2
• /
• pp.203-223
• /
• 2022

We propose a Bayesian approach to cumulative logistic regression model for the ordinal response based on the orthogonal principal components via singular value decomposition considering the multicollinearity among predictors. The advantage of the suggested method is considering dimension reduction and parameter estimation simultaneously. To evaluate the performance of the proposed model we conduct a simulation study with considering a high-dimensional and highly correlated explanatory matrix. Also, we fit the suggested method to a real data concerning sprout- and scab-damaged kernels of wheat and compare it to EM based proportional-odds logistic regression model. Compared to EM based methods, we argue that the proposed model works better for the highly correlated high-dimensional data with providing parameter estimates and provides good predictions.

• Journal of Preventive Medicine and Public Health
• /
• v.37 no.4
• /
• pp.300-305
• /
• 2004

Objectives : The exposure to particulate air pollution during the pregnancy has reported to result in adverse pregnancy outcome such as low birth weight, preterm birth, still birth, and intrauterine growth retardation (IUGR). We aim to assess whether prenatal exposure of particulate matter less than 10 (m in diameter ($PM_{10}$) is associated with preterm birth in Seoul, South Korea. Methods : We included 382,100 women who delivered a singleton at 25-42 weeks of gestation between 1998 and 2000. We calculated the average PM10 exposures for each trimester period and month of pregnancy, from the first to the ninth months, based on the birth date and gestational age. We used three different models to evaluate the effect of air pollution on preterm birth; the logistic regression model, the generalized additive logistic regression model, and the proportional hazard model. Results : The monthly analysis using logistic regression model suggested that the risks of preterm birth increase with PM10 exposure between the sixth and ninth months of pregnancy and the highest risk was observed in the seventh month (adjusted odds ratio=1.07, 95% CI=1.01-1.14). We also found the similar results using generalized additive model. In the proportional hazard model, the adjusted odds ratio for preterm births due to PM10 exposure of third trimester was 1.04 (95% CI=0.96-1.13) and PM10 exposure between the seventh month and ninth months of pregnancy was associated with the preterm births. Conclusions : We found that there were consistent results when we applied the three different models. These findings suggest that air pollution exposure during the third trimester pregnancy has an adverse effect on preterm birth in South Korea.

• Journal of Gastric Cancer
• /
• v.4 no.3
• /
• pp.164-168
• /
• 2004

Purpose: The most feared complication of gastrointestinal tract operations is anastomotic leakage, not only because of the presumed individual surgeon's culpability but also because of the assumption that this event is often fatal. We have experienced 32 cases of anastomotic leakage after elective gastric resection during 8 years. The purpose of this study was to evaluate the result of their treatment. Materials and Methods: We evaluated the records of 1335 patients who had undergone elective gastric resection for an adenocarcinoma of stomach from January 1995 to October 2003 and conducted a retrospective, multivariate analysis. Results: Of the 1335 patients, 32 ($2.4\%$) sustained an anastomotic leakage. Anastomotic leakages usually developed on mean postoperative day $9.1\pm3.2$ (range:$1\∼18$ days).Overall, $31.3\%$ (10/32) of patients who sustained an anastomotic leakage died. The anastomotic leakages were identifed by radiological study or by operative finding at the site of the duodenal stump (20 patients), the esophagojejunostomy (7), the gastroduodenostomy (4), and the gastrojejunostomy (1). Fourteen patients ($43.8\%$) underwent a relaparotomy, a drainage procedure in the main, and 18 patients ($56.3\%$) were treated conservatively. The mortality rates were $42.9\%$ (6/14) and $22.2\%$ (4/18), respectively, but this difference was not statistically significant. A cox's proportional hazard analysis showed that a body-mass Index < 24 kg/m2 (odds ratio 5.55, $95\%$ CI: $0.69\∼44.82$) and non-enteral feeding (odds ratio 18.27, $95\%$ CI 2.22.150.69) were independent factors of mortality due to anastomotic leakage. Conclusion: Our observations show that anastomotic leakage after an elective gastric resection has a high risk of being fatal. Moreover, for a patient with a body-mass index lower than $24\;kg/m^{2}$ and/or non-enteral feeding, an anastomotic leakage after an elective gastric resection has a higher risk of being fatal.

• Journal of Korean Academy of Nursing
• /
• v.36 no.8
• /
• pp.1315-1323
• /
• 2006

Purpose. The present study was intended to compare difference in research variables between delinquent adolescents and student adolescents, and to analyze discriminative factors of delinquent behaviors among Korean adolescents. Methods. The research design of this study was a questionnaire survey. Questionnaires were administered to 2,167 adolescents (1,196 students and 971 delinquents), sampled from 8 middle and high school and 6 juvenile corrective institutions, using the proportional stratified random sampling method. Statistical methods employed were Chi-square, t-test, and logistic regression analysis. Results. The discriminative factors of delinquent behaviors were smoking, alcohol use, other drug use, being sexually abused, viewing time of media violence and pornography. Among these discriminative factors, the factor most strongly associated with delinquency was smoking (odds ratio: 32.32). That is, smoking adolescent has a 32-fold higher possibility of becoming a delinquent adolescent than a non-smoking adolescent. Conclusions. Our findings, that smoking was the strongest discriminative factor of delinquent behavior, suggest that educational strategies to prevent adolescent smoking may reduce the rate of juvenile delinquency. Antismoking educational efforts are therefore urgently needed in South Korea.

• The Korean Journal of Food And Nutrition
• /
• v.32 no.3
• /
• pp.251-258
• /
• 2019

The purpose of this study was to examine the association nutrition education experience in regards to metabolic risk and nutrition intake in Korean adult male using the 2016~2017 Korea National Health and Nutrition Examination Survey as the reference. The study involved a total of 1,978 male subjects aged 40~64 who were classified into the 2 groups based on their nutrition education experience: Educated group (n=88) and non-educated group (n=1,890). The household income and education level of educated group were higher than those of the non-educated group. The two groups showed no significant difference in the level of fasting blood sugar, blood lipid profile including total cholesterol and triglyceride, LDL-cholesterol, and hypertension. Vitamin C intake of the educated group (127.5 mg) was higher compared to the non-educated group (88.2 mg) (p<0.05). The percentage of the subjects utilizing nutritional labels was higher in the educated group. The nutrition education experience was inversely proportional to lower Odds Ratio in hyperLDL-cholesterolemia (OR: 0.47, 95% CI: 0.36, 0.84) and HypoHDL-cholesterolemia (OR: 0.57, 95% CI: 0.37, 0.87). This result indicates that nutritional education can be used as an effective tool to avert chronic diseases and develop healthy eating habits.

• Nutrition Research and Practice
• /
• v.15 no.4
• /
• pp.468-478
• /
• 2021

BACKGROUND/OBJECTIVES: This study aimed to analyze the association between dietary omega-3 fatty acid intake and depression in postmenopausal women using data from the Korea National Health and Nutrition Examination Survey (KNHANES) VI. SUBJECTS/METHODS: The KNHANES is a cross-sectional nationwide health and nutrition survey. Dietary data, including omega-3 fatty acids, were assessed using the 24-h recall method. Depression was evaluated using a survey questionnaire. The association between dietary omega-3 fatty acids and depression was evaluated using multivariate logistic regression analysis. Depression, according to the dietary omega-3 fatty acid intake, was expressed as the odds ratio (OR) with a 95% confidence interval (CI). A total of 4,150 postmenopausal women were included in the analysis. RESULTS: In the fully-adjusted model, the group with the highest dietary omega-3 fatty acid intake significantly showed lower prevalence of depression than the group with the lowest intake (OR, 0.52; 95% CI, 0.33-0.83); a significant linear trend was detected (P for trend = 0.04). According to the dose-response analysis using cubic restricted spline regression, this association was linear and monotonic (P for non-linearity = 0.32). CONCLUSIONS: In this study, the dietary omega-3 fatty acid intake in postmenopausal women was inversely proportional to depression in a dose-response manner. Large cohort studies are needed to verify the causality between omega-3 fatty acids and depression in Korean postmenopausal women.

• Korean Journal of Clinical Pharmacy
• /
• v.32 no.3
• /
• pp.226-237
• /
• 2022

Background: Monoclonal antibodies for the treatment of patients with different types of cancer, such as cetuximab, have been widely used for the past 10 years in oncology. Although drug information package insert contains some representative adverse events which were observed in the clinical trials for drug approval, the overall adverse event patterns on the real-world cetuximab use were less investigated. Also, there have been no published papers that deal with the full spectrums of adverse drug events of cetuximab using national-wide drug safety surveillance systems. Methods: In this study, we detected new adverse event signals of cetuximab in the Korea Adverse Event Reporting System (KAERS) by utilizing proportional reporting ratios, reporting odds ratios, and information components indices. Results: The KAERS database included 869,819 spontaneous adverse event reports, among which 2,116 reports contained cetuximab. We compared the labels of cetuximab among the United States, European Union, Australia, Japan, and Korea to compare the current labeling information and newly detected signals of our study. Some of the signals including hyperkeratosis, tenesmus, folliculitis, esophagitis, neuralgia, disseminated intravascular coagulopathy, and skin/throat tightness were not labeled in the five countries. Conclusion: We identified new signals that were not known at the time of market approval.