臺灣博碩士論文加值系統

心血管疾病在臺灣十大死因中佔了其中的四大項。本篇論文主要的目的是利用一筆來自基隆社區為基礎的資料,研究心血管疾病發病的年紀。所有的參與者在進入研究之前必須是沒有罹患心血管疾病的，並且統一在研究結束時檢查每個人是否罹患心血管疾病。也因為這些篩檢的過程而造成抽樣偏差。因此利用統計的手法將資料視為是左截切與右設限的型態之下進行分析。我們得到了利用無母數所估計出來的發病年紀分佈以及使用Cox模型之下的迴歸分析。除了資料分析之外,我們也利用模擬的手法觀察忽略截切所帶來的影響。

Cardiovascular diseases are among the top 10 causes of death in Taiwan. In the thesis, we aim to study the age-onset for cardiovascular diseases (CVD) using the data collected by the Keelung community-based integrated screening study. All the subjects were free of CVD when they just entered the study and later were examined whether they developed CVD during the study period. Such a selection process, however, produces sampling bias. Therefore we analyze the data by statistical methods which account for the effects of left truncation and right censoring. We obtain nonparametric estimators of the age-onset distribution and then conduct regression analysis under the Cox proportional hazard model. Besides data analysis, we also perform simulations to examine the sensitivity issue when the truncation effect is ignored.

中文摘要 ..............................................................................................................i
英文摘要 .............................................................................................................ii
致謝 ....................................................................................................................iii
Contents ...............................................................................................................iv
List of Tables .......................................................................................................vi
List of Figures ....................................................................................................vii
Chapter 1 Introduction ..................................................................................1
Chapter 2 Preliminary Data Analysis ...........................................................4
2.1 Data description ...........................................................................4
2.2 Data description for different genders..........................................5
2.3 Previous research on KCIS Data ................................................11
Chapter 3 Preliminary Background on Survival Analysis ..........................13
3.1 Notations ....................................................................................13
3.2 Data structure: right censoring ...................................................13
3.3 Data structure: left truncation and right censoring .....................14
Chapter 4 Proposed Nonparametric Analysis .............................................18
4.1 Analysis based on stratified continuous covariates ....................18
4.2 Analysis based on discrete covariates ........................................22
Chapter 5 Proposed Regression Analysis ...................................................26
Chapter 6 Simulation Analysis ...................................................................29
6.1 Data Generation .........................................................................29
6.2 Simulation Results .....................................................................30
6.2.1 The effect of truncation rate .......................................................30
6.2.2 The effect of censoring rate ........................................................31
Chapter 7 Conclusion ..................................................................................33
References ..........................................................................................................34
Appendix ............................................................................................................35

[1] Liu, Y. M., Chen, S. L. S., Yen, A. M. F., &; Chen, H. H. (2013). Individual risk prediction model for incident cardiovascular disease: A Bayesian clinical reasoning approach. International journal of cardiology, 167 (5), 2008-2012.
[2] Kaplan, E. L., &; Meier, P. (1958). Nonparametric estimation from incomplete observations. Journal of the American statistical association, 53 (282), 457-481.
[3] Cox, D. R. (1992). Regression models and life-tables. In Breakthroughs in Statistics (pp. 527-541). Springer New York.
[4] Lynden-Bell, D. (1971). A method of allowing for known observational selection in small samples applied to 3CR quasars. Monthly Notices of the Royal Astronomical Society, 155 (1), 95-118.