• Communications for Statistical Applications and Methods
• /
• v.23 no.4
• /
• pp.335-341
• /
• 2016

Interval censored failure time data often occurs in an observational study where a subject is followed periodically. Instead of observing an exact failure time, two inspection times that include it are made available. Several methods have been suggested to analyze interval censored failure time data (Sun, 2006). In this article, we are concerned with a binary time-varying covariate whose changing time is interval censored. A modified estimating equation is proposed by extending the approach suggested in the presence of a missing covariate. Based on simulation results, the proposed method shows a better performance than other simple imputation methods. ACTG 181 dataset were analyzed as a real example.

• The Korean Journal of Applied Statistics
• /
• v.21 no.4
• /
• pp.659-664
• /
• 2008

Missing continuous covariates are pervasive in the use of generalized linear models for medical data. Multiple imputation is the most common and easy-to-do method of dealing with missing covariate data. However, there are always serious warnings in using this method. There should be concern to make imputed values more proper. In this paper, proper imputation from posterior predictive distribution is developed for implementing with arbitrary priors. We use empirical distribution of the posterior for approximating the posterior predictive distribution, to sample from it. This method is preferable in comparison with a presented imputation method of us which uses a full model to impute missing values using available software. The proposed methods are implemented on glucocorticoid data.

• The Korean Journal of Applied Statistics
• /
• v.16 no.2
• /
• pp.407-426
• /
• 2003

We consider the missing covariates problem in generalized estimating equations(GEE) model. If the covariate is partially missing, GEE can not be calculated. In this paper, we study the performance of 7 imputation methods to handle missing covariates in GEE models, and the properties of GEE estimators are investigated after missing covariates are imputed for ordinal data of repeated measurements. The 7 imputation methods include i) Naive Deletion ii) Sample Average Imputation iii) Row Average Imputation iv) Cross-wave Regression Imputation v) Carry-over Imputation vi) Bayesian Bootstrap vii) Approximate Bayesian Bootstrap. A Monte-Carlo simulation is used to compare the performance of these methods. For the missing mechanism generating the missing data, we assume ignorable nonresponse. Furthermore, we generate missing covariates with or without considering wave nonresp onse patterns.

• Communications for Statistical Applications and Methods
• /
• v.25 no.4
• /
• pp.385-396
• /
• 2018

Competing risks are commonly encountered in biomedical research. Regression models for competing risks data can be developed based on data routinely collected in hospitals or general practices. However, these data sets usually contain the covariate missing values. To overcome this problem, multiple imputation is often used to fit regression models under a MAR assumption. Here, we introduce a multivariate imputation in a chained equations algorithm to deal with competing risks survival data. Using pseudo-observations, we make use of the available outcome information by accommodating the competing risk structure. Lastly, we illustrate the practical advantages of our approach using simulations and two data examples from a coronary artery disease data and hepatocellular carcinoma data.

• Journal of Gastric Cancer
• /
• v.14 no.4
• /
• pp.259-265
• /
• 2014

Purpose: Survival analysis of gastric cancer patients requires knowledge about factors that affect survival time. This paper attempted to analyze the survival of patients with incomplete registered data by using imputation methods. Materials and Methods: Three missing data imputation methods, including regression, expectation maximization algorithm, and multiple imputation (MI) using Monte Carlo Markov Chain methods, were applied to the data of cancer patients referred to the cancer institute at Imam Khomeini Hospital in Tehran in 2003 to 2008. The data included demographic variables, survival times, and censored variable of 471 patients with gastric cancer. After using imputation methods to account for missing covariate data, the data were analyzed using a Cox regression model and the results were compared. Results: The mean patient survival time after diagnosis was $49.1{\pm}4.4$ months. In the complete case analysis, which used information from 100 of the 471 patients, very wide and uninformative confidence intervals were obtained for the chemotherapy and surgery hazard ratios (HRs). However, after imputation, the maximum confidence interval widths for the chemotherapy and surgery HRs were 8.470 and 0.806, respectively. The minimum width corresponded with MI. Furthermore, the minimum Bayesian and Akaike information criteria values correlated with MI (-821.236 and -827.866, respectively). Conclusions: Missing value imputation increased the estimate precision and accuracy. In addition, MI yielded better results when compared with the expectation maximization algorithm and regression simple imputation methods.

• The Korean Journal of Applied Statistics
• /
• v.37 no.2
• /
• pp.225-237
• /
• 2024

In the analysis of competing risks data, some of covariates may not be fully observed for some subjects. In such cases, excluding subjects with missing covariate values from the analysis may result in biased estimates and loss of efficiency. In this paper, we studied multiple imputation and the augmented inverse probability weighting method for regression parameter estimation in the cause-specific proportional hazards model with missing covariates. The performance of estimators obtained from multiple imputation and the augmented inverse probability weighting method is evaluated by simulation studies, which show that those methods perform well. Multiple imputation and the augmented inverse probability weighting method were applied to investigate significant risk factors for the risk of death from breast cancer and from other causes for breast cancer data with missing values for tumor size obtained from the Prostate, Lung, Colorectal, and Ovarian Cancer Screen Trial Study. Under the cause-specific proportional hazards model, the methods show that race, marital status, stage, grade, and tumor size are significant risk factors for breast cancer mortality, and stage has the greatest effect on increasing the risk of breast cancer death. Age at diagnosis and tumor size have significant effects on increasing the risk of other-cause death.

• Communications for Statistical Applications and Methods
• /
• v.16 no.1
• /
• pp.103-113
• /
• 2009

Weighting is a common form of unit nonresponse adjustment in sample surveys where entire questionnaires are missing due to noncontact or refusal to participate. A common approach computes the response weight as the inverse of the response rate within adjustment cells based on covariate information. In this paper, we consider the efficiency and robustness of nonresponse weight adjustment bated on the response propensity and predictive mean. In the simulation study based on 2000 Fishry Census in Korea, the root mean squared errors for assessing the various ways of forming nonresponse adjustment cell s are investigated. The simulation result suggest that the most important feature of variables for inclusion in weighting adjustment is that they are predictive of survey outcomes. Though useful, prediction of the propensity to response is a secondary. Also the result suggest that adjustment cells based on joint classification by the response propensity and predictor of the outcomes is productive.

• Journal of Environmental Health Sciences
• /
• v.33 no.4
• /
• pp.227-234
• /
• 2007

Recent epidemiologic studies show that gestational exposure to air pollution adversely affects pregnancy outcomes including low birth weight in preform birth. In this study, we evaluated the effect of air pollutants on LBW (low birth weight) on firstborn fetus throughout the gestational period using the birth cohort between 1999 and 2003 in Seoul. Using birth cohort data from the National Statistics Office of Korea we identified 288,346 firstborn births (excluded missing data on lack of information for birth weight and discordance between residential and certificated address from a total of 316,451) during 1999 to 2003 with complete covariate (gender, parity, date of birth, gestational age, parental age and educational level, maternal occupation etc.) and maternal residential history data. Our subjects were defined as more than 37 weeks and less than 44 weeks of completed gestation and we identified 5,457 persons (1.89%) by low birth weight (<2.5 kg) in this study. Using logistic regression, we estimated the risk of mean (entire pregnancy and trimester period) air pollution concentrations for CO, $O_3,\;PM_{10},\;NO_2\;and\;SO_2$. In terms of trimester-specific exposure, we found that some air pollutants exposure in each trimester would increase the risk for LBW. Results also showed that the effect size of air pollutants exposure during the first and third trimester is higher than during the second trimester. In all trimester, the estimated risk of LBW was 1.831 (95% CI=1.573-2.132) with unit increase for CO, 1.139 (95% CI=1.107-1.172) for 50, and 1.009 (95% CI=1.001-1.017) for $O_3$. Our results suggest that exposure during the gestation period to relatively low levels of some air pollutants may be associated with a reduction in birth weight on first-born fetus. These findings implicate the effective risk management strategies should be applied to minimize the public health impacts for pregnant women.

• Journal of Gastric Cancer
• /
• v.17 no.2
• /
• pp.132-144
• /
• 2017

Purpose: To identify baseline prognostic factors for survival in patients with disease progression, during or after chemotherapy for the treatment of advanced gastric or gastroesophageal junction (GEJ) cancer. Materials and Methods: We pooled data from patients randomized between 2009 and 2012 in 2 phase III, global double-blind studies of ramucirumab for the treatment of advanced gastric or GEJ adenocarcinoma following disease progression on first-line platinum- and/or fluoropyrimidine-containing therapy (REGARD and RAINBOW). Forty-one key baseline clinical and laboratory factors common in both studies were examined. Model building started with covariate screening using univariate Cox models (significance level=0.05). A stepwise multivariable Cox model identified the final prognostic factors (entry+exit significance level=0.01). Cox models were stratified by treatment and geographic region. The process was repeated to identify baseline prognostic quality of life (QoL) parameters. Results: Of 1,020 randomized patients, 953 (93%) patients without any missing covariates were included in the analysis. We identified 12 independent prognostic factors of poor survival: 1) peritoneal metastases; 2) Eastern Cooperative Oncology Group (ECOG) performance score 1; 3) the presence of a primary tumor; 4) time to progression since prior therapy <6 months; 5) poor/unknown tumor differentiation; abnormally low blood levels of 6) albumin, 7) sodium, and/or 8) lymphocytes; and abnormally high blood levels of 9) neutrophils, 10) aspartate aminotransferase (AST), 11) alkaline phosphatase (ALP), and/or 12) lactate dehydrogenase (LDH). Factors were used to devise a 4-tier prognostic index (median overall survival [OS] by risk [months]: high=3.4, moderate=6.4, medium=9.9, and low=14.5; Harrell's C-index=0.66; 95% confidence interval [CI], 0.64-0.68). Addition of QoL to the model identified patient-reported appetite loss as an independent prognostic factor. Conclusions: The identified prognostic factors and the reported prognostic index may help clinical decision-making, patient stratification, and planning of future clinical studies.