• The Korean Journal of Applied Statistics
• /
• v.24 no.6
• /
• pp.1045-1054
• /
• 2011

During clinical trials a researcher is frequently able to observe a disease symptom in a subject as well as a severity score for those who experienced a symptom after a fixed length of treatment. The traditional method to evaluate a decreasing trend in proportion, when there is an intrinsic order in the treatment groups (for example control and two or more treatment groups) is a Cochran-Armitage test, while the method to evaluate a decreasing trend in continuous non-normal data is a Jonckheere-Tersptra test. The Cochran-Armitage test emphasizes the dichotomous data of symptom occurrence and the Jonckheere-Tersptra test emphasizes the continuous non-normal data of severity symptom scores. In this paper we propose new test statistics that consider the combined evidence from a symptom occurrence and disease severity score. We illustrate these methods with example data of schizophrenic inpatients that demonstrated antipsychotic-drug induced constipation. A small-scale simulation is conducted to compare the new trend tests with other trend tests.

• The Korean Journal of Applied Statistics
• /
• v.21 no.1
• /
• pp.1-17
• /
• 2008

The ordinary least squares regression method of Haseman and Elston(1972) is most widely used in genetic linkage studies for continuous traits of sib pairs. Kruglyak and Lander(1995) suggested a statistic which appears to be a nonparametric counterpart to the Haseman and Elston(1972)'s regression method, but in fact these two methods are quite different. In this paper the relationships between these two methods are described and will be compared by simulation studies. One of the characteristics of the sib-pair linkage study is that the explanatory variable has only three different values and thus dependent variable is heavily replicated in each value of the explanatory variable. We propose a weighted least squares regression method which is more appropriate to this situation and the efficiency of the weighted regression in genetic linkage study was explored with normal and non-normal simulated continuous traits data. Simulation studies demonstrated that the weighted regression is more powerful than other tests.

• Communications for Statistical Applications and Methods
• /
• v.15 no.6
• /
• pp.925-937
• /
• 2008

In this study, we considered the clustering analysis for stock return traded in the stock market. Most of financial time-series data, for instance, stock price and exchange rate have conditional heterogeneous variability depending on time, and, hence, are not properly applied to the autoregressive moving-average(ARMA) model with assumption of constant variance. Moreover, the variability is font and center for stock investors as well as academic researchers. So, this paper focuses on the generalized autoregressive conditional heteroscedastic(GARCH) model which is known as a solution for capturing the conditional variance(or volatility). We define the metrics for similarity of unconditional volatility and for homogeneity of model structure, and, then, evaluate the performances of the metrics. In real application, we do clustering analysis in terms of volatility and structure with stock return of the 11 Korean companies measured for the latest three years.

• The Korean Journal of Applied Statistics
• /
• v.22 no.4
• /
• pp.793-804
• /
• 2009

The hypotheses of difference, ratio and odds ratio between two proportions are used for the non-inferiority trial. The approximate unconditional test suggested by Kang and Chen (2000) based on difference and ratio have the potential problem against the failure rate. When the sample size is small, the type I errors of the asymptotic test using the normal approximation suggested by Chen et al. (2000) tends to exceed the nominal level. Therefore, we propose the approximate unconditional test based on odds ratio and compare the test with the asymptotic test. And we compare the three hypotheses used in the approximate unconditional tests of two proportions with respect to the type I errors and power.

• Journal of Korean Neurosurgical Society
• /
• v.52 no.3
• /
• pp.193-199
• /
• 2012

Objective : The objective of study is to evaluate the incidence of leptomeningeal carcinomatosis (LMC) in breast cancer patients with parenchymal brain metastases (PBM) and clinical risk factors for the development of LMC. Methods : We retrospectively analyzed 27 patients who had undergone surgical resection (SR) and 156 patients with whole brain radiation therapy (WBRT) as an initial treatment for their PBM from breast cancer in our institution and compared the difference of incidence of LMC according to clinical factors. The diagnosis of LMC was made by cerebrospinal fluid cytology and/or magnetic resonance imaging. Results : A total of 27 patients (14%) in the study population developed LMC at a median of 6.0 months (range, 1.0-50). Ten of 27 patients (37%) developed LMC after SR, whereas 17 of 156 (11%) patients who received WBRT were diagnosed with LMC after the index procedure. The incidence of LMC was significantly higher in the SR group compared with the WBRT group and the hazard ratio was 2.95 (95% confidence interval; 1.33-6.54, p<0.01). Three additional factors were identified in the multivariable analysis : the younger age group (<40 years old), the progressing systemic disease showed significantly increased incidence of LMC, whereas the adjuvant chemotherapy reduce the incidence. Conclusion : There is an increased risk of LMC after SR for PBM from breast cancer compared with WBRT. The young age (<40) and systemic burden of cancer in terms of progressing systemic disease without adjuvant chemotherapy could be additional risk factors for the development of LMC.

• The Korean Journal of Applied Statistics
• /
• v.22 no.5
• /
• pp.1059-1072
• /
• 2009

In the analysis of microarray data with a small number of arrays, the most important task is the detection of differentially expressed genes by a significance test. For this purpose, one needs to construct a null distribution based on a large number of genes and one of the best way for constructing the null distribution for a small number of arrays is by means of permutation methods. In this paper we propose simple test statistics and permutation methods that are appropriate in constructing the null distribution. In a simulation study, we compare the null distributions generated by the proposed test statistics and permutation methods with the previous ones. With an example microarray data, differentially expressed genes are determined by applying these methods.

• The Journal of Korean Society for School & Community Health Education
• /
• v.11 no.1
• /
• pp.67-78
• /
• 2010

Background & Objectives: Current trend is that funding agencies require investigators to share their data with others. However, there is limited guidance how to access and utilize the shared data. We sought to determine what common data sharing practices in U.S.A. are, what data-related to adolescent health are freely available, and how we deal with the large dataset adopting the complex study design. Methods: The study included only research data-related to adolescent health which was collected in USA and unlimitedly accessible through the internet. Only the raw data, not aggregated, was considered for the study. Major keywords for web search were "adolescent", "children", "health", and "school". Results: Current approaches for public health data sharing lacked of common standards and varied largely due to the data's complex nature, large size, local expertise and internal procedures. Some common data sharing practices are unlimited access, formal screened access, restricted access, and informal exclusive access. The Inter-University Consortium for Political and Social Research and the Center for Disease Control and Prevention were the best data depository. "Data on the net" was search engine for the website providing data freely available. Six datasets related to adolescent health freely available were identified. The importance and methods of incorporating complex research design into analysis was discussed. Conclusion: There have been various attempts to standardize process for open access and open data using the information technology concept. However, it may not be easy for researchers to adapt themselves to this high technology. Therefore, guidance provided by this study may help researchers enhance the accessibility to and the utilization of the open source data.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.23
• /
• pp.10335-10337
• /
• 2015

Background: Gastric cancer is considered as the second most prevalent cancer in Iran. The present research sought to identify high risk clusters of gastric cancer with mapping using space-time scan statistics. Materials and Methods: The present research is of descriptive type. The required data were gathered from the registered cancer reports of Cancer Control Office in the Center for Non Communicable Disease of the Ministry of Health (MOH). The data were extracted at province level in the time span of 2004-9. Sat-Scan software was used to analyse the data and to identify high risk clusters. ArcGIS10 was utilized to map the distribution of gastric cancer and to demonstrate high risk clusters. Results: The most likely clusters were found in Ardabil, Gilan, Zanjan, East-Azerbaijan, Qazvin, West-Azerbaijan, Kurdistan, Hamadan, Tehran and Mazandaran between 2007 and 2009. It was statistically significant at the p-value below 0.05. Conclusions: High risk regions included Northern, West-North and central provinces, particularly Ardabil, Kurdistan, Mazandaran and Gilan. More screening tests are suggested to be conducted in high risk regions along with more frequent epidemiological studies to enact gastric cancer prevention programs.

• Asian Pacific Journal of Cancer Prevention
• /
• v.13 no.6
• /
• pp.2639-2642
• /
• 2012

Background: The patterns of gastric cancer recurrence vary across societies. We designed the current study in an attempt to evaluate and reveal the outbreak of the recurrence patterns of gastric cancer and also prediction of time to recurrence and its effected factors in Iran. Materials and Methods: This research was performed from March 2003 to February 2007. Demographic characteristics, clinical and pathological diagnosis and classification including pathologic stage, tumor grade, tumor site and tumor size in of patients with GC recurrent were collected from patients' data files. To evaluate of factors affected on the relapse of the GC patients, gender, age at diagnosis, treatment type and Hgb were included in the research. Data were analyzed using Kaplan-Meier and logistic regression models. Results: After treatment, 82 patients suffered recurrence, 42, 33 and 17 by the ends of first, second and third years. The mean ( SD) and median ( IQR) time to recurrence in patients with GC were 25.5 (20.6-30.1) and 21.5 (15.6-27.1) months, respectively. The results of multivariate analysis logistic regression showed that only pathologic stage, tumor grade and tumor site significantly affected the recurrence. Conclusions: We found that pathologic stage, tumor grade and tumor site significantly affect on the recurrence of GC which has a high positive prognostic value and might be functional for better follow-up and selecting the patients at risk. We also showed time to recurrence to be an important factor for follow-up of patients.

• Asian Pacific Journal of Cancer Prevention
• /
• v.13 no.6
• /
• pp.2857-2860
• /
• 2012

Pancreatic cancer is usually detected late and has a high mortality rate. Since little is known about this cancer in Malaysia, a review of all cases admitted to Universiti Sains Malaysia Hospital was conducted to identify the epidemiological distribution and assess survival. A list of pancreatic cancer patients in 2001-2008 was obtained from the Hospital Record Department. Only cases confirmed by radio-imaging or histo-pathology examination were included. We excluded those with incomplete medical records. Kaplan-Meier and Cox proportional hazard approaches were used for data analysis. Only 56 cases were included with a mean (SD) age of 49.6 (16.0) years, with 60.7% males and 82.1% of Malay ethnicity. Previous history included cholelithiasis in 23.2%, diabetes mellitus in 16.1%, previous laparotomy in 10.7%, chronic pancreatitis in 7.1%, alcohol drinking in 5.4% and positive family history in 3.6%. The common presenting history included 67.9% loss of appetite, 66.1% loss of weight, 58.9% jaundice and 46.4% abdominal pain. Tumour staging was: 21.5% stage l, 17.8% stage ll, 3.6% stage lll and 57.1% stage lV. The median (95% CI) survival time was 3.4 (0.5, 6.3) months and significant prognostic factors were duration of symptoms (HR 0.97; 95% CI: 0.95, 0.99; p value 0.013), ascites (HR 2.64; 95% CI: 1.28, 5.44; p value 0.008) and Whipple surgery (HR 4.20; 95% CI: 2.27, 7.76; p value <0.001). The history of presenting complaints was short and the majority presented at late stages of the disease, thus the median survival time was very poor.