<b>Background:</b> 'Competing risks' are events that either preclude or alter the probability of experiencing the primary study outcome(s). Many standard survival models fail to account for competing risks, introducing an unknown level of bias in their measures of absolute and relative risk. Individuals with type 2 diabetes mellitus (T2DM) and albuminuria are at increased risk of multiple competing causes of mortality, including cardiovascular disease (CVD), cancer and renal disease, yet studies to date have not implemented competing risks methodology. <b>Aim:</b> Using albuminuria in T2DM as a case study, this Thesis set out to quantify differences between standard- and competing-risks-adjusted survival analysis estimates of absolute and relative risk for the outcomes of cardiovascular and cancer mortality. <b>Methods:</b> 86,962 patients aged &ge;35 years with T2DM present on or before 2005 were identified in the Clinical Practice Research Datalink. To quantify differences in measures of absolute risk, cumulative risk estimates for cardiovascular and cancer mortality from standard survival analysis methods (Kaplan-Meier estimator) were compared to those from competing-risks-adjusted methods (cumulative incidence competing risk estimator). Cumulative risk estimates were stratified by patient albuminuria level (normoalbuminuria vs albuminuria). To quantify differences in measures of relative risk, estimates for the effect of albuminuria on the relative hazards of cardiovascular and cancer mortality were compared between standard cause-specific hazard (CSH) models (Cox-proportional-hazards regression), competing risk CSH models (unstratified Lunn-McNeil model), and competing risk subdistribution hazard (SDH) models (Fine-Gray model). <b>Results:</b> Patients with albuminuria, compared to those with normoalbuminuria, were older (p&LT;0.001), had higher systolic blood pressure (p&LT;0.001), had worse glycaemic control (p&LT;0.001), and were more likely to be current or ex-smokers (p&LT;0.001). Over the course of nine years of follow-up 22,512 patients died; 8,800 from CVD, 5,239 from cancer, and 8,473 from other causes. Median follow-up was 7.7 years. In patients with normoalbuminuria, nine-year standard and competing-risks-adjusted cumulative risk estimates for cardiovascular mortality were 11.1% (95% confidence interval (CI): 10.8-11.5%) and 10.2% (95% CI: 9.9-10.5%), respectively. For cancer mortality, these figures were 8.0% (95% CI: 7.7-8.3%) and 7.2% (95% CI: 6.9-7.5%). In patients with albuminuria, standard and competing-risks-adjusted estimates for cardiovascular mortality were 21.8% (95% CI: 20.9-22.7%) and 18.5% (95% CI: 17.8-19.3%), respectively. For cancer mortality, these figures were 10.7% (95% CI: 10.0-11.5%) and 8.6% (8.1-9.2%). For the effect of albuminuria on cardiovascular mortality, hazard ratios from multivariable standard CSH, competing risks CSH, and subdistribution hazard ratios from competing risks SDH models were 1.75 (95% CI: 1.63-1.87), 1.75 (95% CI: 1.64-1.87), and 1.58 (95% CI: 1.48-1.69), respectively. For the effect of albuminuria on cancer mortality, these values were 1.27 (95% CI: 1.16-1.39), 1.28 (95% CI: 1.17-1.40), and 1.11 (95% CI: 1.01-1.21). <b>Conclusions:</b> When evaluating measures of absolute risk, differences between standard and competing-risks-adjusted methods were small in absolute terms, but large in relative terms. For the investigation of epidemiological relationships using relative hazards models, standard survival analysis methods produced near-identical risk estimates to the CSH competing risks methods for the clinical associations evaluated in this Thesis. For the evaluation of risk prediction using relative hazards models, CSH models produced consistently higher risk estimates than SDH models, and their use may lead to over-estimation of the predictive effect of albuminuria on either outcome. Where outcomes are less common (like cancer) CSH models provide poor estimates of risk prediction, and SDH models should be used. This research demonstrates that differences can be present between risk estimates derived using CSH and SDH methods, and that the two are not necessarily interchangeable. Moreover, such differences may be present in other clinical areas.

The current study consists of three major parts. Statistical modeling, the connection between statistical modeling and cluster analysis, and proposing new methods to cluster time dependent information. First, we perform a statistical modeling of the Carbon Dioxide (CO2) emission in South Korea in order to identify the attributable variables including interaction effects. One of the hot issues in the earth in 21st century is Global warming which is caused by the marriage between atmospheric temperature and CO2 in the atmosphere. When we confront this global problem, we first need to verify what causes the problem then we can find out how to solve the problem. Thereby, we find and rank the attributable variables and their interactions based on their semipartial correlation and compare our findings with the results from the United States and European Union. This comparison shows that the number one contributing variable in South Korea and the United States is Liquid Fuels while it is the number 8 ranked in EU. This comparison provides the evidence to support regional policies and not global, to control CO2 in an optimal level in our atmosphere. Second, we study regional behavior of the atmospheric CO2 in the United States. Utilizing the longitudinal transitional modeling scheme, we calculate transitional probabilities based on effects from five end-use sectors that produce most of the CO2 in our atmosphere, that is, the commercial sector, electric power sector, industrial sector, residential sector, and the transportation sector. Then, using those transitional probabilities we perform a hierarchical clustering procedure to classify the regions with similar characteristics based on nine US climate regions. This study suggests that our elected officials can proceed to legislate regional policies by end-use sectors in order to maintain the optimal level of the atmospheric CO2 which is required by global consensus. Third, we propose new methods to cluster time dependent information. It is almost impossible to find data that are not time dependent among floods of information that we have nowadays, and it needs not to emphasize the importance of data mining of the time dependent information. The first method we propose is called “Lag Target Time Series Clustering (LTTC)” which identifies actual level of time dependencies among clustering objects. The second method we propose is the “Multi-Factor Time Series Clustering (MFTC)” which allows us to consider the distance in multi-dimensional space by including multiple information at a time. The last method we propose is the “Multi-Level Time Series Clustering (MLTC)” which is especially important when you have short term varying time series responses to cluster. That is, we extract only pure lag effect from LTTC. The new methods that we propose give excellent results when applied to time dependent clustering. Finally, we develop appropriate algorithm driven by the analytical structure of the proposed methods to cluster financial information of the ten business sectors of the N.Y. Stock Exchange. We used in our clustering scheme 497 stocks that constitute the S&P 500 stocks. We illustrated the usefulness of the subject study by structuring diversified financial portfolio.

This thesis examines cancer and mortality rates among people diagnosed with hepatitis B (HBV) and C (HCV) infection in New South Wales (NSW) from 1990 through 2002, by linking hepatitis notifications with the NSW Central Cancer Registry (CCR) and National Death Index. Of the 39101 HBV, 75834 HCV and 2604 HBV/HCV co-infection notifications included 1052, 1761 and 85 were linked to cancer notifications and 1233, 4008 and 186 were linked to death notifications respectively. Of 2072 hepatocellular carcinoma (HCC) notifications to the CCR 323, 267 and 85 were linked to HBV, HCV and HBV/HCV co-infection notifications. Incidence of HCC was 6.5, 4.0 and 5.9 per 1000 person years for HBV, HCV and HBV/HCV co-infected groups. Risk of HCC in those diagnosed with hepatitis was 20 to 30 times greater than the standard population. There was a marginally statistically significant increased risk of immunoproliferative malignancies associated with HCV infection (SIR=5.6 95% CI 1.8 ???17.5). Risk of death for those with hepatitis was significantly greater, 1.5 to 5 fold, than the general population with the greatest risk among those with HBV/HCV co-infection. The primary cause of HBV deaths was liver related, particularly HCC, whereas in the HCV groups drug related deaths were most frequent. Among people with HCV, risk of dying from drug related causes was significantly greater than from liver related causes (p=0.012), with the greatest increased risk in females age 15- 24 years (SMR 56.9, 95%CI 39.2???79.9). Median age at diagnosis of HCC varied markedly by country of birth and hepatitis group: HBV 66, 63 and 57years ; HCV 51, 68 and 71 years; unlinked 69, 70 and 64 years for Australian, European, and Asian-born groups, respectively (P<0.0001 for all groups). While the risk of cancer, particularly HCC, is elevated among people with HBV and HCV infection, the absolute risk remains low. Young people with HCV face a higher mortality risk from continued drug use than from liver damage related to their HCV infection. The influence of IDU in the epidemiology of HCC in New South Wales was possibly reflected in the varying distributions of age and country of birth.

This thesis examines cancer and mortality rates among people diagnosed with hepatitis B (HBV) and C (HCV) infection in New South Wales (NSW) from 1990 through 2002, by linking hepatitis notifications with the NSW Central Cancer Registry (CCR) and National Death Index. Of the 39101 HBV, 75834 HCV and 2604 HBV/HCV co-infection notifications included 1052, 1761 and 85 were linked to cancer notifications and 1233, 4008 and 186 were linked to death notifications respectively. Of 2072 hepatocellular carcinoma (HCC) notifications to the CCR 323, 267 and 85 were linked to HBV, HCV and HBV/HCV co-infection notifications. Incidence of HCC was 6.5, 4.0 and 5.9 per 1000 person years for HBV, HCV and HBV/HCV co-infected groups. Risk of HCC in those diagnosed with hepatitis was 20 to 30 times greater than the standard population. There was a marginally statistically significant increased risk of immunoproliferative malignancies associated with HCV infection (SIR=5.6 95% CI 1.8 ???17.5). Risk of death for those with hepatitis was significantly greater, 1.5 to 5 fold, than the general population with the greatest risk among those with HBV/HCV co-infection. The primary cause of HBV deaths was liver related, particularly HCC, whereas in the HCV groups drug related deaths were most frequent. Among people with HCV, risk of dying from drug related causes was significantly greater than from liver related causes (p=0.012), with the greatest increased risk in females age 15- 24 years (SMR 56.9, 95%CI 39.2???79.9). Median age at diagnosis of HCC varied markedly by country of birth and hepatitis group: HBV 66, 63 and 57years ; HCV 51, 68 and 71 years; unlinked 69, 70 and 64 years for Australian, European, and Asian-born groups, respectively (P<0.0001 for all groups). While the risk of cancer, particularly HCC, is elevated among people with HBV and HCV infection, the absolute risk remains low. Young people with HCV face a higher mortality risk from continued drug use than from liver damage related to their HCV infection. The influence of IDU in the epidemiology of HCC in New South Wales was possibly reflected in the varying distributions of age and country of birth.

Orientador: Tirza Aidar / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Filosofia e Ciencias Humanas / Made available in DSpace on 2018-08-06T06:10:50Z (GMT). No. of bitstreams: 1 Belon_AnaPaula_M.pdf: 999877 bytes, checksum: 35afba21b9b0ad79c323d894c75e5ec6 (MD5) Previous issue date: 2006 / Resumo: O cenário demográfico que se delineia no Estado de São Paulo é caracterizado pelo aumento envelhecimento relativo populacional e da participação relativa e das taxas de mortalidade por neoplasias malignas entre as causas de óbito. A estreita associação entre mortes por neoplasias malignas e a população idosa reforça a importância deste estudo, que apresenta como proposta investigar a relação entre as condições de vida dos idosos e a mortalidade por neoplasias de colo de útero, mama feminina e próstata para o Estado de São Paulo no ano de 2000. Parte-se do pressuposto que as desigualdades socioeconômicas se expressam nos diferenciais da mortalidade por neoplasias entre idosos e seu comportamento ao longo do tempo. Resgata-se as dimensões socioeconômicas e demográficas da mortalidade, numa tentativa de não se restringir à simples mensuração da desigualdade em saúde. Elege-se, como variáveis socioeconômicas para compor perfil socioeconômico dos idosos, os anos de estudo e rendimento domiciliar per capita, tendo como categorias de referência o analfabetismo funcional e o rendimento igual ou superior a 5 s.m. per capita. Para tanto, o Estado de São Paulo é dividido em Direções Regionais de Saúde (DIR) e a população idosa em grupos etários qüinqüenais e por sexo. Os anos censitários, que auxiliam a compreensão da evolução temporal, são 1980, 1991 e 2000. Através de análises de correlação e graus de dispersão, a dissertação aponta como resultados que: (1) ocorre um aumento mais significativo das taxas específicas de mortalidade por neoplasias entre idosos com idades mais avançadas no decorrer dos anos; (2) quanto maior a participação relativa do analfabetismo funcional entre os responsáveis pelo domicílio, menor são os riscos de morrer por neoplasias malignas; (3) quanto maior a proporção de domicílios com rendimento per capita igual ou superior a 5 s.m., maiores são as taxas específicas de mortalidade; (4) a localização e distribuição dos centros de saúde de alta complexidade, segundo as DIR¿s influem na magnitude das taxas; (5) as neoplasias de mama feminina e próstata apresentam maiores índices de correlação entre as taxas e as variáveis socioeconômicas, sendo que o comportamento de colo de útero seria mais aleatório / Abstract: There is a demographic scenery for the State of São Paulo (Brazil) characterized by population ageing and an increasing rate of death, among this population, caused by malignant neoplasms. Based on these findings, this study intends to investigate the relation between socioeconomics and demographic pointers and mortality by malignant neoplasms ¿ uterine cervical, feminine breast and prostate ¿ among the aged population of the State of São Paulo and its health regional services during the year of 2000. Presuming that the socioeconomics inequalities are expressed in the mortality rates by malignant neoplasms among aged people, it was elected as variables to compose the socioeconomic profiles, schooling and per capita domicile income. The reference categories are determined as functional illiteracy and the income of 5 minimal salaries or above per capita. The State of São Paulo is divided by the ¿regional health services¿ (DIR) and the aged population by sex in 5-aged groups. The census years which helps to understand the time evolution are 1980, 1991 and 2000. Through descriptive analysis as well as linear models adjusts, the results suggest that: (1) there is a significative increase in the mortality rates by malignant neoplasms among the eldest and this tendency does not present a homogeneity aspect among the DIR¿s; (2) the rate of mortality due to feminine breast and prostata cancers is, unexpectedly, higher in the more developed regions; (3) in areas with health centers of high complexity for cancer treatment, the same tendency occurs, i.e. the highest levels of deaths as a consequence of neoplasms in aged population were observed. / Mestrado / Saude e Morbi-mortalidade / Mestre em Demografia

Lung cancer has been considered one of the leading causes of deaths while cancer re- mains the second most common cause of deaths in the USA. Understanding the behavior of a disease over time could yield important information to make decisions about the disease. Statistical models could provide crucial clues and help to make a decision about the dis- ease, budget allocation, evaluation, and implement prevention. Longitudinal trend analysis of the diseases helps to understand long term effects and nature. Cholesterol level is one of the most contributing risk factors for Coronary Heart Disease. Studying cholesterol statistically helps to know more about its nature and provides crucial information to mitigate its effectiveness in diagnosing its impact to public health. In our study, we have analyzed lung cancer mortality in the USA based on age at death, period at death, and birth cohort to investigate its nature in longitudinal effects. The attempt has been made to estimate mortality rate based on age for different age groups and to find the relative risk of mortality due to period effect and relative risk due to birth cohort for lung cancer in the United States. Our statistical analysis and modeling are based on the data obtained from Surveillance Epidemiology and End Results (SEER) program of the United States. We have also investigated the probabilistic behavior of average cholesterol level based on gender and ethnicity. The study reveals significant differences with respect to the distribution they follow and their basic inferences which could be beneficial to draw conclusions in various ways in addressing related issues. At the same time, the change of cholesterol level over time for an individual might be a good source to study the association of cholesterol level, coronary heart disease and their effects on age. The cholesterol data is obtained from inter-university Consortium for Political and Social Research and National Health and Nutrition Examination Survey (NHANS) of the United States. Understanding the average change in total serum cholesterol level over time as people get older could be vital to explore it. We have studied the longitudinal behavior of the association of sex and time with cholesterol level. It is observed that age, sex, and time have an individual effect and can impact differently upon collective considerations. Their adverse effect in increasing cholesterol level could promote to worsen the cholesterol re- lated issues and hence heart related diseases. We believe our study pivots knowing more about target population of cholesterol level and helps to have the useful inference about cholesterol levels for public health. Finally, we also analyzed the average cholesterol data using a functional data analysis approach to understand its nature and effect on age. Since functional data analysis approach presents more flexibility in modeling, it could provide more insight in studying cholesterol level.

by Kai-Hang Choi. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1998. / Includes bibliographical references (leaves 217-232). / Abstract also in Chinese. / ABSTRACT --- p.i / ACKNOWLEDGMENT --- p.iv / TABLE OF CONTENTS --- p.v / LIST OF FIGURES --- p.viii / LIST OF TABLES --- p.xiii / Chapter CHAPTER I --- INTRODUCTION --- p.1 / Chapter 1.1 --- Background --- p.1 / Chapter 1.2 --- Role of GIS in Health Studies --- p.4 / Chapter 1.3 --- Research Objectives --- p.5 / Chapter 1.4 --- Organization of the Thesis --- p.6 / Chapter CHAPTER II --- LITERATURE REVIEW --- p.8 / Chapter 2.1 --- Introduction --- p.8 / Chapter 2.2 --- Human cancer --- p.8 / Chapter 2.3 --- Environment and Cancer --- p.10 / Chapter 2.4 --- Cancer Etiology and Epidemiology --- p.13 / Chapter 2.5 --- Observational Cancer Epidemiology --- p.15 / Chapter 2.6 --- Geography of Cancer --- p.17 / Chapter 2.7 --- Geographical Epidemiology of Cancer --- p.19 / Chapter 2.7.1 --- Geographical Variation in Cancer Occurrence --- p.21 / Chapter 2.7.1.1 --- Cancer Mapping --- p.24 / Chapter 2.7.1.2 --- Spatial Autocorrelation --- p.26 / Chapter 2.7.2 --- Identifying Causal Association --- p.29 / Chapter 2.7.3 --- Environmental Factors of Cancer --- p.31 / Chapter 2.8 --- Geographical Information Systems --- p.40 / Chapter 2.9 --- GIS and Health --- p.41 / Chapter 2.9.1 --- GIS Applications in Health Planning --- p.42 / Chapter 2.9.2 --- GIS Applications in Health Research --- p.43 / Chapter 2.10 --- Cancer Studies with GIS --- p.45 / Chapter 2.11 --- Conclusion --- p.47 / Chapter CHAPTER III --- THE STUDY AREA AND RESEARCH METHODOLOGY --- p.49 / Chapter 3.1 --- Introduction --- p.49 / Chapter 3.2 --- Disease Transition in Hong Kong --- p.49 / Chapter 3.3 --- Cancer in Contemporary Hong Kong --- p.52 / Chapter 3.3.1 --- Trends of Cancer Mortality and Incidence --- p.52 / Chapter 3.3.2 --- The Common Types of Cancer --- p.55 / Chapter 3.3.3 --- Geographical Variation of Cancer in Hong Kong --- p.58 / Chapter 3.4 --- The Research --- p.61 / Chapter 3.4.1 --- Cartographic Analysis --- p.62 / Chapter 3.4.2 --- Statistical Analyses --- p.63 / Chapter 3.4.3 --- Cancer Variables --- p.67 / Chapter 3.4.4 --- Environmental Variables --- p.70 / Chapter 3.5 --- Conclusion --- p.71 / Chapter CHAPTER IV --- DATABASE CONSTRUCTION --- p.73 / Chapter 4.1 --- Introduction --- p.73 / Chapter 4.2 --- Data Collection --- p.73 / Chapter 4.2.1 --- Base Maps --- p.73 / Chapter 4.2.2 --- Cancer Data --- p.74 / Chapter 4.2.3 --- Socio-demographic Data --- p.75 / Chapter 4.2.4 --- Air Pollution --- p.76 / Chapter 4.2.5 --- ELF EMFs --- p.77 / Chapter 4.3 --- Data Input --- p.77 / Chapter 4.3.1 --- Spatial Data --- p.77 / Chapter 4.3.1.1 --- Base Maps --- p.78 / Chapter 4.3.1.2 --- Point Data --- p.78 / Chapter 4.3.1.3 --- Line Data --- p.79 / Chapter 4.3.2 --- Attribute Data --- p.79 / Chapter 4.4 --- Data Editing and Conversions --- p.80 / Chapter 4.4.1 --- Spatial Data --- p.80 / Chapter 4.4.1.1 --- Standard Coverage Editing Procedures --- p.80 / Chapter 4.4.1.2 --- Specific Coverage Editing Procedures --- p.81 / Chapter 4.4.2 --- Attribute Data --- p.83 / Chapter 4.4.2.1 --- Cancer Rates --- p.83 / Chapter 4.4.2.2 --- Socio-economic Status --- p.85 / Chapter 4.5 --- Data Pre-processing and Manipulation --- p.86 / Chapter 4.5.1 --- Socio-economic Variables --- p.86 / Chapter 4.5.1.1 --- Interpretation of Factor Scores --- p.97 / Chapter 4.5.2 --- Compromised Traffic Index --- p.99 / Chapter 4.5.3 --- ELFEMFs --- p.104 / Chapter 4.6 --- Conclusion --- p.106 / Chapter CHAPTER V --- RESULTS AND DISCUSSIONS --- p.111 / Chapter 5.1 --- Introduction --- p.111 / Chapter 5.2 --- Geographical Analysis of Cancer Patterns --- p.111 / Chapter 5.2.1 --- Results --- p.112 / Chapter 5.2.1.1 --- Total Cancer --- p.113 / Chapter 5.2.1.2 --- Cancer of the Female Breast --- p.118 / Chapter 5.2.1.3 --- Cancer of the Cervix Uteri (Cervical Cancer) --- p.121 / Chapter 5.2.1.4 --- Cancer of the Colon and Rectum (Colorectal Cancer) --- p.124 / Chapter 5.2.1.5 --- Cancer of the Stomach (Gastric Cancer) --- p.129 / Chapter 5.2.1.6 --- Leukaemia --- p.129 / Chapter 5.2.1.7 --- Cancer of the Liver --- p.134 / Chapter 5.2.1.8 --- Cancer of the Lung --- p.143 / Chapter 5.2.1.9 --- Cancer of the Nasopharynx (NPC) --- p.149 / Chapter 5.2.1.10 --- Cancer of the Oesophagus --- p.154 / Chapter 5.3 --- Correlation among Cancer Variables --- p.160 / Chapter 5.3.1 --- Correlation among Cancer types --- p.160 / Chapter 5.3.2 --- Temporal Correlation among Cancers --- p.168 / Chapter 5.3.3 --- Correlation between Cancer Mortality and Incidence --- p.170 / Chapter 5.4 --- Correlation between Cancer and Environmental Variables --- p.172 / Chapter 5.4.1 --- Results --- p.174 / Chapter 5.5 --- Weighted Stepwise Regression Modeling --- p.182 / Chapter 5.5.1 --- Results --- p.183 / Chapter 5.5.1.1 --- Total Cancer --- p.184 / Chapter 5.5.1.2 --- Cancer of the Female Breast --- p.186 / Chapter 5.5.1.3 --- Cancer of the Cervix Uteri (Cervical Cancer) --- p.188 / Chapter 5.5.1.4 --- Cancer of the Colon and Rectum --- p.189 / Chapter 5.5.1.5 --- Cancer of the Stomach (Gastric Cancer) --- p.191 / Chapter 5.5.1.6 --- Leukaemia --- p.193 / Chapter 5.5.1.7 --- Cancer of the Liver --- p.195 / Chapter 5.5.1.8 --- Cancer of the Lung --- p.197 / Chapter 5.5.1.9 --- Cancer of the Nasopharynx (NPC) --- p.199 / Chapter 5.5.1.10 --- Cancer of the Oesophagus --- p.201 / Chapter 5.6 --- Interpretations of Results --- p.203 / Chapter CHAPTER VI --- CONCLUSION --- p.207 / Chapter 6.1 --- Summary of Findings --- p.207 / Chapter 6.1.1 --- Summary on Geographical Analysis of Cancer Patterns --- p.207 / Chapter 6.1.2 --- Summary on Statistical Analysis of Cancer Variables --- p.209 / Chapter 6.1.3 --- Summary on Associations between Cancers and Environment --- p.211 / Chapter 6.2 --- Research Limitations --- p.212 / Chapter 6.3 --- Implications for Future Studies --- p.215 / BIBLIOGRAPHY --- p.217 / APPENDICES --- p.233 / Appendix I Community Map of hong Kong --- p.234 / Appendix II List of Communities and their Components --- p.236 / Appendix III Tertiary Planning Units (TPUs) - Community Conversion Lists --- p.240 / Appendix IV BASIC Program for Calculating Moran and Geary Indices --- p.244

by Siu-Fai Leung. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1998. / Includes bibliographical references (leaves 80-81). / Abstract also in Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Mortality and Socioeconomic Status --- p.1 / Chapter 1.2 --- Research Plan and Difficulties Encountered in the Study --- p.4 / Chapter 2 --- Imputation and File Merging --- p.8 / Chapter 2.1 --- Structure and Contents of Data Sets --- p.8 / Chapter 2.2 --- Imputation of Missing Values --- p.14 / Chapter 2.3 --- Merging Data Sets --- p.22 / Chapter 2.3.1 --- Merging Death Data and Census Data --- p.22 / Chapter 2.3.2 --- Merging Two Census Data Sets --- p.29 / Chapter 2.3.3 --- Final Data Set Used in Modeling --- p.31 / Chapter 3 --- Modeling and Estimation --- p.33 / Chapter 3.1 --- Discrete-Time Hazard Function Analysis --- p.33 / Chapter 3.1.1 --- The Hazard Function --- p.34 / Chapter 3.1.2 --- Logistic Regression --- p.36 / Chapter 3.2 --- Application of Discrete-Time Hazard Model on the Death Data Set --- p.37 / Chapter 3.2.1 --- Preparing the Person-Period Data Set --- p.38 / Chapter 3.2.2 --- Modeling the Person-Period Data Set --- p.41 / Chapter 3.3 --- Combining Results from different imputed data sets --- p.47 / Chapter 3.4 --- Estimation of Cell Probabilities --- p.51 / Chapter 4 --- Model Adequacy Checking --- p.52 / Chapter 4.1 --- The Definition of Residuals in Multiple Imputation --- p.52 / Chapter 4.2 --- Residual Analysis of The Cancer Mortality Model --- p.59 / Chapter 5 --- Conclusion --- p.63 / Chapter 5.1 --- The Cancer Mortality --- p.63 / Chapter 5.2 --- Competing Risk --- p.68 / Chapter 5.3 --- Discussion --- p.72 / Appendix A: Coding Description of District --- p.75 / Appendix B: Results of the Heart Diseases Mortality Model --- p.76 / Bibliography --- p.80

臺灣的癌症發生率及死亡率有連年增加的趨勢，研究指出原因可能與環境中的污染物質有關，檢測可能的污染源附近是否存在癌症群聚(Cluster)，將有助於未來的癌症防治。在空間統計(Spatial Statistics)有不少方法可用於檢測群聚現象，其中用來檢測某個特定位置周圍是否發生群聚的方法被稱為焦點檢定(Focused Test)，本文介紹及評估常用的焦點檢定方法，並使用較佳方法探討臺灣地區疑似污染源的地區。 首先本文使用電腦模擬，在不同情境假設下比較焦點檢定方法的檢定力(Power)，例如研究區域大小、群聚形狀等不同的情境，以判斷檢定方法之間的優劣。最後本文分析臺灣鄉鎮市(Township)層級癌症死亡資料，應用焦點檢定方法分析石門核一廠、恆春核三廠及麥寮六輕周圍的癌症死亡率，檢定結果發現核一廠及麥寮六輕附近有較高的癌症死亡率。 / The cancer incidence and mortality rate in Taiwan have been increasing over the past 30 years. Previous studies indicate that the pollution sources, especially for those creating air pollution and excess radiation, are one of the potential causes for the increment. Correctly, detecting the location of possible sources of contaminants can help for cancer prevention. In spatial statistics, focused test can be used to determine if the intensity rate are higher around a possible pollution source. We will introduce and evaluate frequently used focused tests and apply them in Taiwan. First we use computer simulation to compare the power of focused tests in different scenarios, such as study region and cluster shape. Next, we apply the focused tests to Taiwan cancer mortality data, in order to decide if the cancer mortality rates are higher around Chinshan nuclear power plant, Maanshan nuclear power plant, and Mailiao sixth naphtha cracker. The results show that the cancer mortality rates around Chinshan nuclear power plant and Mailiao sixth naphtha cracker are significantly higher.

by Chiu Yuk Lan. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1998. / Includes bibliographical references (leaves 127-133). / Abstract also in Chinese. / TABLE OF CONTENTS / ABSTRACT (ENGLISH) --- p.a / ABSTRACT (CHINESE) --- p.b / ACKNOWLEDGEMENTS --- p.iv / Chapter CHAPTER 1 --- INTRODUCTION --- p.1 / Chapter 1.1 --- Cancer in Food-service Workers --- p.1 / Chapter 1.2 --- Carcinogenicity of Cooking Fumes --- p.1 / Chapter 1.3 --- High Risk of Lung Cancer in Chinese Women --- p.2 / Chapter 1.4 --- Why do We Conduct This Study? --- p.3 / Chapter 1.5 --- Implication of This Study --- p.4 / Chapter 1.6 --- What Types of Cancer were Included in This Study? --- p.4 / Chapter 1.7 --- Aims and Hypothesis of This Study --- p.5 / Chapter 1.8. --- Outline of the Thesis --- p.5 / Chapter CHAPTER 2 --- LITERATURE REVIEW --- p.8 / Chapter 2.1. --- Occupational Epidemiological Studies --- p.8 / Chapter 2.1.1 --- Studies of occupation and cancer occurrence based on routine records --- p.8 / Chapter 2.1.2 --- Retrospective cohort studies among food service workers --- p.21 / Chapter 2.1.3 --- Case-control studies --- p.27 / Chapter 2.1.4 --- Case reports --- p.29 / Chapter 2.1.5 --- Summary --- p.29 / Chapter 2.2. --- Mutagens and Carcinogens in Cooking Fumes --- p.39 / Chapter 2.2.1 --- Mutagens and carcinogens in cooking fumes --- p.40 / Chapter 2.2.2 --- Summary --- p.42 / Chapter CHAPTER 3 --- METHODS --- p.44 / Chapter 3.1 --- Study Design --- p.44 / Chapter 3.2 --- Study Population and Subjects --- p.46 / Chapter 3.3 --- Reference Population --- p.48 / Chapter 3.4 --- Sample Size Estimation --- p.48 / Chapter 3.5 --- Data Sources and Data Collection --- p.49 / Chapter 3.6 --- Data Processing --- p.53 / Chapter 3.7 --- Data Analyses --- p.54 / Chapter 3.7.1 --- Standardised proportional mortality ratio (SPMR) --- p.54 / Chapter 3.7.2 --- Adjusted' SPMRs --- p.56 / Chapter 3.7.3 --- Mortality odds ratio (MOR) --- p.58 / Chapter 3.8. --- Exploring if Smoking could be a Confounding Factor --- p.62 / Chapter CHAPTER 4 --- RESULTS --- p.64 / Chapter 4.1 --- Characteristics of the Food-service Workers --- p.64 / Chapter 4.2 --- Cancer Mortality Patterns of Food-service Workers --- p.69 / Chapter 4.3 --- Adjusted SPMRs --- p.72 / Chapter 4.4 --- Mortality Odds Ratios (MORs) --- p.76 / Chapter 4.5 --- Mortality Odds Ratios Using Multiply Reference Diseases --- p.77 / Chapter 4.6. --- Comparing SPMRs with MORs --- p.82 / Chapter 4.7. --- Internal Comparison --- p.83 / Chapter 4.8 --- Summary of Results --- p.90 / Chapter 4.9. --- Survey on Smoking and Drinking Prevalence among Current Food-service Workers --- p.92 / Chapter 4.9.1 --- Smoking habit --- p.92 / Chapter 4.9.2 --- Drinking habit --- p.94 / Chapter CHAPTER 5 --- DISCUSSION OF FINDINGS --- p.95 / Chapter 5.1 --- Outcomes for This Study --- p.95 / Chapter 5.1.2 --- Cancer risks for the kitchen workers --- p.96 / Chapter 5.1.3 --- Cancer risks for the outside kitchen workers --- p.102 / Chapter 5.2 --- Limitations of the Methods Adopted in the Present study --- p.107 / Chapter 5.2.1 --- Standardised proportional mortality ratio (SPMR) --- p.107 / Chapter 5.2.2 --- Morality odds ratio (MOR) --- p.109 / Chapter 5.3 --- Bias and Control --- p.111 / Chapter 5.3.1 --- Selection bias --- p.111 / Chapter 5.3.2 --- Information bias --- p.113 / Chapter 5.3.3 --- Confounding --- p.116 / Chapter 5.4 --- Implications from the Results of the Present Study --- p.117 / Chapter 5.5 --- Conclusion --- p.119 / APPENDIX --- p.121 / Appendix 1 --- p.121 / Appendix 2 --- p.123 / Appendix 3 --- p.124 / Appendix 4 --- p.125 / REFERENCES --- p.127