| Liu Yuanyuan,Li Changping,Hu Liangping,Analysis of regression model of survival data——overview of survival data and its statistical analysis methods[J].SICHUAN MENTAL HEALTH,2020,33(1):21-26 |
| Analysis of regression model of survival data——overview of survival data and its statistical analysis methods |
DOI：10.11886/scjsws20200106002 |
| English keywords:Survival data Censored data Survival analysis Survival function Hazard function Regression analysis |
| Fund projects:国家自然科学基金项目（项目名称：贝叶斯生存分析方法在肝细胞癌肝移植患者预后预测中的应用研究，项目编号：81803333） |
| Author Name | Affiliation | Postcode | | Liu Yuanyuan | School of Public Health, Tianjin Medical University, Tianjin 300070, China | 300070 | | Li Changping | School of Public Health, Tianjin Medical University, Tianjin 300070, China
Specialty Committee of Clinical Scientific Research Statistics of World Federation of Chinese Medicine Societies, Beijing 100029, China | 100029 | | Hu Liangping | Specialty Committee of Clinical Scientific Research Statistics of World Federation of Chinese Medicine Societies, Beijing 100029, China
Graduate School, Academy of Military Sciences PLA China, Beijing 100850, China | 100850 |
|
| Hits: |
| Download times: |
| English abstract: |
| The purpose of this paper was to provide a comprehensive introduction to the characteristics of survival data and its commonly used statistical analysis methods. Survival data had the following four prominent features: ①survival data had both survival outcomes and survival times; ②the survival times might contain censored data or truncated data; ③the distribution of survival times usually did not follow the normal distribution, and the survival times were often distributed in the exponential distribution, Weibull distribution, lognormal distribution, and so on; ④the factors affecting the survival times were complicated and difficult to control. The statistical analysis methods of survival data involves three categories: statistical description, difference analysis and regression analysis. The statistical description methods mainly included Kaplan-Meier estimation method and Life table estimation method. The difference analyses mainly included log-rank test and Wilcoxon test. The regression analyses mainly included Cox's proportional and non-proportional hazards regression models and parametric regression models. In the statistical analysis of the survival data, it was necessary to select the statistical analysis methods reasonably, in order to comprehensively and deeply reveal the inherent changes of the survival data. |
| View Full Text
View/Add Comment Download reader |
| Close |
|
|
|