• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.1
• /
• pp.483-488
• /
• 2014

Aim: This study employed public use National Health and Nutrition Examination Survey (NHANES III) data to investigate the association between urinary cadmium (UDPSI) and all cause, all cancer and prostate cancer mortalities in men. Patients and Methods: NHANES III household adult, laboratory and mortality data were merged. The sampling weight used was WTPFEX6, with SDPPSU6 applied for the probability sampling unit and SDPSTRA6 to designate the strata for the survey analysis. Results: For prostate cancer death, the significant univariates were UDPSI, age, weight, and drinking. Under multivariate logistic regression, the significant covariates were age and weight. For all cause mortality in men, the significant covariates were UDPSI, age, and poverty income ratio. For all cancer mortality in men, the significant covariates were UDPSI, age, black and Mexican race. Conclusions: UDPSI was a predictor of all cause and all cancer mortalities in men as well as prostate cancer mortality.

• Asian Pacific Journal of Cancer Prevention
• /
• v.14 no.4
• /
• pp.2259-2263
• /
• 2013

Background: Public use National Health and Nutrition Examination Survey (NHANES III) and NHANES III linked mortality data were here applied to investigate the association between health insurance coverage and all cause and all cancer mortality in adults. Patients and Methods: NHANES III household adult, laboratory and mortality data were merged. Only patients examined in the mobile examination center (MEC) were included in this study. The sampling weight employed was WTPFEX6, SDPPSU6 being used for the probability sampling unit and SDPSTRA6 to designate the strata for the survey analysis. All cause and all cancer mortalities were used as binary outcomes. The effect of health insurance coverage status on all cause and all cancer mortalities were analyzed with potential socioeconomic, behavioral and health status confounders. Results: There were 2398 sample persons included in this study. The mean age was 40 years and the mean (S.E.) follow up was 171.85 (3.12) person months from the MEC examination. For all cause mortality, the odds ratios (significant p-values) of the covariates were: age, 1.0095 (0.000); no health insurance coverage (using subjects with health insurance), 1.71 (0.092); black race (using non-Hispanic white subjects as the reference group) 1.43, (0.083); Mexican-Americans, 0.60 (0.089); DMPPIR, 0.82, (0.000); and drinking hard liquor, 1.014 (0.007). For all cancer mortality, the odds ratio (significant p-values) of the covariates were: age, 1.0072 (0.00); no health insurance coverage, using with health coverage as the reference group, 2.91 (0.002); black race, using non-Hispanic whites as the reference group, 1.64 (0.047); Mexican Americans, 0.33 (0.008) and smoking, 1.017 (0.118). Conclusion: There was a 70% increase in risk of all cause death and almost 300% of all cancer death for people without any health insurance coverage.

• Asian Pacific Journal of Cancer Prevention
• /
• v.14 no.5
• /
• pp.3105-3108
• /
• 2013

Background: This study used National Health and Nutrition Examination Survey III to study the relationship between blood lead concentration and all cause, all cancer and lung cancer mortality in adults. Patients and Methods: Public use National Health and Nutrition Examination Survey (NHANES III) data were used. NHANES III uses stratified, multistage probabilistic methods to sample nationally representative samples. Household adult, laboratory and mortality data were merged. Sample persons who were available to be examined in aMobile Examination Center (MEC) were included in this study. Specialized survey analysis software was used. Results: A total of 3,482 sample participants with complete information for all variables were included in this analysis. For all cause death, the odds ratios (S.E.) for statistically significant variables were body mass index, 1.03 (1.01-1.06); 1.01 (1.01-1.01); blood lead concentration, 1.05 (1.01-1.08); poverty income ratio, 0.823 (0.76-0.89); and drinking hard liquor, 1.01 (1.00-1.02). For all cancer mortality, the odds ratios (S.E.) of the statistically signigicant variables were: age, 1.01 (1.01-1.01); blood lead concentration, 1.07 (1.04-1.12), black race, using non-Hispanic white as reference, 1.69 (1.12-2.56); and smoking, 1.02 (1.01-1.04). For lung cancer mortality, the odds ratios (S.E.) of the statistically significant variables were: age, 1.01(1.01-1.01); blood lead concentration, 1.09 (1.05-1.13); Mexican Americans, using non-Hispanic white as refrence, 0.33 (0.129-0.850); other races, 1.80 (0.53-6.18); and smoking, 1.03 (1.02-1.05). Conclusion: Blood lead concentration correlated with all cause, all cancer, and lung cancer mortality in adults.

• Asian Pacific Journal of Cancer Prevention
• /
• v.14 no.5
• /
• pp.2707-2710
• /
• 2013

Background: This study used pre-hepatitis A vaccination era data in U.S. to study the relationship between serum hepatitis A antibody positivity with pancreas cancer mortality in adults. Patients and Methods: Public use National Health and Nutrition Examination Survey (NHANES III) data were employed. NHANES III uses complex probabilistic methods to sample nationally representative samples. Household adult laboratory and mortality data were merged. Sample persons who were available to be examined in the Mobile Examination Center (MEC) were included in this study. All results were obtained by using specialized survey software taking into account the primary sampling unit and stratification variables and the weights assigned to the sample persons examined in the MEC. Thus they are representative of the U.S. population. Results: The mean risk (95%CI) of death in the study population for pancreas cancer was 0.0014 (-0.000069 -.0029); their mean age (95%CI) at the mobile examination center (MXPAXTMR) was 473.43 (463.85-482.10); the follow up in months from their medical examination (permth_exm) was 170.12 (164.17-176.07). The odds ratios (S.E.) of the statistically significant univariables were: age, 1.007 (1.005-1.009); serum anti-hepatitis antibody status, 0.038 (0.004-0.376); and drinking hard liquor, 1.014 (1.004-1.023). The coefficients (S.E.) of the statistically significant variables after multivariate analysis were 0.006 (0.002-0.010) for age and -2.528 (-4.945--0.111) for serum anti-hepatitis A antibody negativity (using serum anti-hepatitis A antibody positivity as a reference). Conclusion: Serum hepatitis A antibody positivity correlates with higher pancreas cancer mortality in adults.

• Proceedings of the KSCN Conference
• /
• 1998.05a
• /
• pp.63-63
• /
• 1998

• Communications for Statistical Applications and Methods
• /
• v.10 no.2
• /
• pp.553-566
• /
• 2003

In many cases, the measurement error variances may be functions of the unknown true values or related covariates. This paper considers design-based estimators of the parameters of these variance functions based on the within-unit sample variances. This paper devotes to: (1) define an error scale factor $\delta$; (2) develop estimators of the parameters of the linear measurement error variance function of the true values under large-sample and small-error conditions; (3) use propensity methods to adjust survey weights to account for possible selection effects at the replicate level. The proposed methods are applied to medical examination data from the U.S. Third National Health and Nutrition Examination Survey (NHANES III).

• 한국데이터정보과학회:학술대회논문집
• /
• 2003.05a
• /
• pp.1-14
• /
• 2003

In man cases, the measurement error variances may be functions of the unknown true values or related covariate. In some cases, the measurement error variances increase in proportion to the value of predictor. This paper develops estimators of the parameters of a linear measurement error variance function under stratified multistage random sampling design and additional conditions. Also, this paper evaluates and compares the power of an asymptotically unbiased test with that of an asymptotically biased test. The proposed method are applied to blood sample measurements from the U.S. Third National Health and Nutrition Examination Survey(NHANES III)

• The Journal of the Korean dental association
• /
• v.44 no.10 s.449
• /
• pp.658-670
• /
• 2006

1980년대 후반기부터 치주질환과 뇌경색(ischemic stroke)자료의 연관성을 모색하는 시도가 있어왔다. 이번 연구의 목적은 치주질환과 뇌경색 유무와의 어떤 관계가 있는지를 60세 이상의 노인을 대상으로 조사, 통계 분석하였다. 자료는 미국의 총 국민조사 격인 The Third Nation Health and Nutrition Examination Survey (NHANES III)를 이용하였다. 이번 연구에서 unadjusted logistic model 통계법을 이용하여 치아 상실수와 뇌경색 경험이 통계학적으로 유의한 수치의 상관성이 있음을 알게 되었다. 또한 나이와 흡연유무를 고려, 조정한 후 multiple logistic model 통계법으로 잔존치아가 적을수록 더욱 뇌경색에 걸릴 확률이 높음을 보였다. 그러나 두 질병에 동시에 선택된 중요한 위험인자 (risk factor)를 모두 고려, 조정 한 후에는 통계학적인 유의성을 찾지 못했다. 치은퇴축, 치주낭 깊이, 치석, 탐침시 출혈과 뇌경색 경험은 각각의 비교법에서 약간의 상관성을 보이나, 모든 통계법을 통해 일괄된 결과를 얻을 수는 없었다.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.1
• /
• pp.245-254
• /
• 2016

Many surveys provide categorical data and there may be one or more missing categories. We describe a nonignorable nonresponse model for the analysis of two-way contingency tables from small areas. There are both item and unit nonresponse. One approach to analyze these data is to construct several tables corresponding to missing categories. We describe a hierarchical Bayesian model to analyze two-way categorical data from different areas. This allows a "borrowing of strength" of the data from larger areas to improve the reliability in the estimates of the model parameters corresponding to the small areas. Also we use a nonignorable nonresponse model with Bayesian uncertainty analysis by placing priors in nonidentifiable parameters instead of a sensitivity analysis for nonidentifiable parameters. We use the griddy Gibbs sampler to fit our models and compute DIC and BPP for model diagnostics. We illustrate our method using data from NHANES III data on thirteen states to obtain the finite population proportions.

• Journal of the Korean Data and Information Science Society
• /
• v.26 no.6
• /
• pp.1547-1555
• /
• 2015

We study the problem of nonignorable nonresponse in a two-way contingency table and there may be one or two missing categories. We describe a nonignorable nonresponse model for the analysis of two-way categorical table. One approach to analyze these data is to construct several tables (one complete and the others incomplete). There are nonidentifiable parameters in incomplete tables. We describe a hierarchical Bayesian model to analyze two-way categorical data. We use a nonignorable nonresponse model with Bayesian uncertainty analysis by placing priors in nonidentifiable parameters instead of a sensitivity analysis for nonidentifiable parameters. To reduce the effects of nonidentifiable parameters, we project the parameters to a lower dimensional space and we allow the reduced set of parameters to share a common distribution. We use the griddy Gibbs sampler to fit our models and compute DIC and BPP for model diagnostics. We illustrate our method using data from NHANES III data to obtain the finite population proportions.