• The Korean Journal of Applied Statistics
• /
• v.30 no.2
• /
• pp.203-213
• /
• 2017

We study estimation and inference of joint conditional distributions of bivariate longitudinal outcomes using regression models and copulas. We consider a class of time-varying transformation models and combine the two marginal models using Gaussian copulas to estimate the joint models. Our models and estimation method can be applied in many situations where the conditional mean-based models are inadequate. Gaussian copulas combined with time-varying transformation models may allow convenient and easy-to-interpret modeling for the joint conditional distributions for bivariate longitudinal data. We apply our method to an epidemiological study of repeatedly measured bivariate cholesterol data.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.3
• /
• pp.689-700
• /
• 2016

We study estimation and inference of the joint conditional distributions of bivariate longitudinal outcomes using regression models and copulas. For the estimation of marginal models we consider a class of time-varying transformation models and combine the two marginal models using nonparametric empirical copulas. Regression parameters in the transformation model can be obtained as the solution of estimating equations and our models and estimation method can be applied in many situations where the conditional mean-based models are not good enough. Nonparametric copulas combined with time-varying transformation models may allow quite flexible modeling for the joint conditional distributions for bivariate longitudinal data. We apply our method to an epidemiological study of repeatedly measured bivariate cholesterol data.

• Communications for Statistical Applications and Methods
• /
• v.20 no.4
• /
• pp.283-290
• /
• 2013

Subjects can experience two types of recurrent events in a longitudinal study. In addition, there may exist intermittent dropouts that results in repeated observation gaps during which no recurrent events are observed. Therefore, theses periods are regarded as non-risk status. In this paper, we consider a special case where information on the observation gap is incomplete, that is, the termination time of observation gap is not available while the starting time is known. For a statistical inference, incomplete termination time is incorporated in terms of interval-censored data and estimated with two approaches. A shared frailty effect is also employed for the association between two recurrent events. An EM algorithm is applied to recover unknown termination times as well as frailty effect. We apply the suggested method to young drivers' convictions data with several suspensions.

• Journal of Engineering Education Research
• /
• v.23 no.1
• /
• pp.26-36
• /
• 2020

This study aims to investigate the effect of college choice factors on withdrawal intention of first-year undergraduates in STEM majors. The data for this study were collected from a national sample of 4-year institutions that participated in Collegiate Longitudinal Study funded by the National Research Foundation of Korea during 2013-2014. The data included 1,484 students in STEM majors from 30 institutions, and this study used bivariate and multivariate analytic methods to answer the research questions. This study found that while the internal dimensions of college choice showed a negative effect on the intention to dropout, external dimensions of college choice such as college environment, financial accessibility, admission possibility had a positive effect on the intention. The findings indicate that it is necessary to pay attention to college choice for enrollment management to prevent first-year undergraduates in STEM majors from dropping out.

• Journal of Periodontal and Implant Science
• /
• v.47 no.2
• /
• pp.96-105
• /
• 2017

Purpose: The National Health Insurance Service-National Sample Cohort and medical checkup data from 2002 to 2013 were used to evaluate the association between periodontal surgery for the treatment of periodontitis (PSTP) and vasculogenic erectile dysfunction (VED). Methods: Bivariate and multivariate logistic regression analyses were applied to a longitudinal retrospective database to assess the association between PSTP and VED while adjusting for the potential confounding effects of sociodemographic factors (age, household income, insurance status, health status, residence area, and smoking status) and comorbidities (diabetes mellitus, angina pectoris, cerebral infarction, and myocardial infarction). Results: Among the 7,148 PSTP within the 268,296 recruited subjects, the overall prevalence of VED in PSTP was 1.43% (n=102). The bivariate analysis showed that VED was significantly related to PSTP (odds ratio [OR], 1.99; 95% confidence interval [CI], 1.38-2.06; P<0.001), and this was confirmed in the multivariate analysis after adjusting for sociodemographic factors and comorbidities (OR, 1.29; 95% CI, 1.06-1.58; P=0.002). Conclusions: Subjects with a history of periodontal flap surgery had a significantly higher risk of VED, after adjusting for potential confounding factors. Further studies are required to identify the key mechanisms underlying the association between severe periodontal disease and VED.

• Korean Journal of Adult Nursing
• /
• v.26 no.6
• /
• pp.621-629
• /
• 2014

Purpose: The purpose of this study was to investigate the nutritional support, gastric residual volume, and nutritional status of the intensive care unit (ICU) patients on enteral feeding. Methods: A descriptive longitudinal design was used to collect 5 day data on enteral nutrition of 52 ICU patients in an university hospital. Nutritional support was calculated with actual caloric intake compared to individual caloric requirement. Residual volumes were measured prior to routine feedings, and the serum albumin levels and the total lymphocyte counts were checked to evaluate nutritional status. The data were analyzed using one group repeated measures ANOVA, paired t-test, and Spearman's bivariate correlation analysis. Results: The subjects received their first enteral feeding on the $5.75^{th}$ day of ICU admission. The mean nutritional support rate was 49.1% of the requirement, however prescription rate and support rate were increased as time goes by. Gastric residual volumes were less than 10 cc in 95% cases. A significant negative correlation was found between nutritional support and nutritional status. Conclusion: The nutritional support for ICU patient was low compared to the requirement, and their nutritional status was worse than at the time of ICU admission. Further studies are necessary to develop nursing interventions for improving nutritional support for ICU patients.

• Journal of Preventive Medicine and Public Health
• /
• v.37 no.2
• /
• pp.174-181
• /
• 2004

Objective : To examine the association between dietary factors and cognitive impairment in older Korean women living in the community. Methods : Wave 2 data, from the Suwon Longitudinal Aging Study (SLAS), of 365 women aged 65 years or over, were used. The Korean version of the 30-point Mini-Mental State Examination (MMSE-K) was used to assess cognitive impairment (score$\geq$19). Dietary habits and frequencies of food group consumption were also examined. Results : A total of 67 women (18.4%) were found to be cognitively impaired. In bivariate analyses, nondietary factors, such as age, marital status, education, income, self-rated health, depression, emotional support, social activity, exercise, and dietary factors, such as self-rated nutritional status, frequency of beans and bean products and milk and dairy products consumption were associated with cognitive impairment. In the multivariate analysis, a higher frequency ($\geq$1 vs. <1 time/day) of beans and bean products (OR=0.48, 95% CI: 0.23-0.99) and milk and dairy products (OR=0.25, 95% CI: 0.10-0.61) consumption was inversely associated with cognitive impairment, after adjustment for non-dietary factors. Conclusion : These results suggest that dietary factors may play a significant role in cognitive impairment of older Korean women.

• Korean Journal of Social Welfare
• /
• v.67 no.3
• /
• pp.281-303
• /
• 2015

This study examined the interdependent retirement decision between husbands and wives. Since most of retirees in Korea retired by involuntary reasons such as health problems, it purposes to test whether the involuntary retirement of husbands(wives) influences the retirement decision of wives(husbands). Using data from 4th Korean Longitudinal Study of Ageing Panel, it conducted probit and multivariate probit analysis on 1,038 dual-earner couples aged 51~76 to test interdependence of couples' retirement. The results showed that husbands who have retired wife were more likely to retire. It also proved that involuntary retirement of spouse is a significant predictor of early retirement decision, though the direction of the effect is reversed by gender. Wives were more likely to retire when husband were retired involuntarily, while husbands were more likely to stay in labor market. These findings helped to understand the retirement decision of aged couples in Korea and brought suggestion on labor market and care policies for aged couples.

• Journal of Korean Society of Occupational and Environmental Hygiene
• /
• v.16 no.4
• /
• pp.324-333
• /
• 2006

This study was designed to investigate the effect of bone demineralization and tibia lead on blood lead in retired lead workers. Two hundred thirty five(126 females and 109 males) retired lead workers who worked in 4 different lead factories and 101 non-occupationally lead exposed subjects(51 females and 51 males) were recruited from March 2004 to October 2004. Bone mineral density(BMD) was measured at left calcaneous bone area by broadband ultrasound attenuation(BUA) method with QUS-2(Metra Biosystems Inc, USA). The BUA value transformed into T-score by WHO standard conversion criteria. Tibia bone lead was measured for skeletal bone lead with K-xray fluorescence(K-XRF) and blood lead was analyzed with flameless atomic spectrophotometer. Hemoglobin, hematocrit, serum calcium and iron were also analyzed. In addition, information for smoking and drinking status and basic personal data such as age, gender and lead exposure were also collected using questionnaire inquiry. Blood lead was correlated with tibia lead (r=0.664) and these two variables were negatively correlated with BMD in bivariate analysis. BMD showed significant main effect on the change of blood lead independent to tibia lead without any effect modification of age or gender; the one T-score unit decrease of mineral bone density made $0.43{\mu}g/dl$ increase of blood lead. On the other hand, tibia lead showed effect modification with gender on blood lead; the slope of tibia lead on blood lead in male was steeper than in female and crossed at around zero of tibia lead. In the multiple regression analysis of blood lead and tibia lead on BMD after adjustment of related covariates, only blood lead showed statistically significant effect on BMD. This study confirmed that BMD and blood lead were significantly associated. To verify the causal association of BMD on blood lead and vice versa, further longitudinal studies are needed.

• Journal of Korean Society of Occupational and Environmental Hygiene
• /
• v.20 no.1
• /
• pp.19-28
• /
• 2010

This study was designed to investigate the effect of increased blood and tibia lead on the change of bone mineral density in retired male lead workers. One hundred nine retired male lead workers who worked in 4 different lead industries and 51 nonoccupationally lead exposed male subjects were recruited from March 2004 to October 2004. Bone mineral density(BMD) was measured by broadband ultrasound attenuation(BUA) at left calcaneous bone area with broadband ultrasound attenuation method of QUS-2(Metra Biosystems Inc, USA). Tibia bone lead was measured for skeletal bone lead with K-xray fluorescence(K-XRF) and blood lead was analyzed with flameless atomic spectrophotometer. Hemoglobin, hematocrit, serum calcium and iron were also analyzed. In addition, information for smoking and drinking status and basic personal data such as age and lead exposure were also collected using questionnaire inquiry. Blood lead was correlated with tibia lead (r=0.711) and these two variables were negatively correlated with BUA in bivariate analysis. BUA and tibia lead showed significant main effects on the change of blood lead after adjusting covariates. The effect modification by the level of BMD (low: lower than the median of BUA and high: higher than the median of BUA) was observed between the association of tibia lead and blood lead after adjustment of covariates. The subjects who had higher BMD seemed to have lower blood lead by the increase of tibia lead than those of lower BMD. In the multiple regression analysis of blood lead and tibia lead on BUA after adjustment of covariates, only blood lead showed statistically significant effect on BUA. This study confirmed that BMD and blood lead were significantly associated. To verify the causal association of BMD on blood lead and vice versa, further longitudinal studies are needed.