竞争风险分析方法研究及其在林县普通人群营养干预试验随访研究中的应用
发布时间：2018-07-15 17:29
【摘要】：目的:系统深入的评价目前常用的竞争风险分析方法,探究竞争事件对目标结局事件发生风险的影响,比较不同分析方法之间的优劣,提出竞争风险分析的一般策略,为其在实际研究中的应用提供理论依据。材料和方法:在林县普通人群营养干预试验随访研究中,利用目前常用的竞争风险分析方法深入全面的评价新鲜水果摄入与食管癌长期死亡风险之间的关联,探究竞争事件在随访期间对食管癌长期死亡风险的影响,估计食管癌的基线死亡风险,同时评价不同竞争风险分析方法之间的优劣,提出竞争风险分析的一般策略。结果:1、当存在竞争风险偏倚时,相对风险函数和相对风险比可以充分的评价竞争事件对目标结局事件发生风险的影响效应;虚拟观测值法可以用于目标结局事件基线发生风险的估计;与原因别风险回归分析的结果相比,使用累积风险回归分析可以更准确的估计所研究因素与目标结局事件之间的关联。2、在林县普通人群营养干预试验随访队列人群中,原因别风险回归分析的结果显示:与不摄入新鲜水果受试人群相比,每周摄入新鲜水果1次以上的受试者食管癌的长期死亡风险可能降低,且这种保护性的效应并不随着新鲜水果摄入频率的增加而增加(P_Trend=0.444);累积风险回归分析的结果则表明:新鲜水果的摄入可能不能降低食管癌的长期死亡风险,提示竞争风险偏倚的存在夸大了新鲜水果摄入与食管癌长期死亡风险之间的关联。竞争风险分析的结果显示:竞争事件对食管癌死亡风险的影响效应随着随访时间的延长而增大,尤其是心脑血管疾病的死亡,且在不同新鲜水果摄入频率组中竞争风险效应的表现形式不同。进一步分析发现:在该队列人群中食管癌的基线死亡风险为1.39,校正新鲜水果的摄入频率后的食管癌基线死亡风险为1.62,提示新鲜水果的摄入肯能不能降低该队列人群中食管癌的长期死亡风险。结论:1、在医学随访研究中,推荐使用累积风险回归分析来评价所研究因素与疾病之间的关联,同时使用相对风险函数和/或相对危险比来描述竞争事件对疾病发生风险的影响,提高研究结果的可解读性。2、新鲜水果的摄入可能不能降低林县营养干预试验随访队列人群中食管癌的长期死亡风险。
[Abstract]:Objective: to evaluate the current commonly used methods of competitive risk analysis, to explore the impact of competition events on the risk of target outcome events, to compare the advantages and disadvantages of different analysis methods, and to put forward the general strategy of competitive risk analysis. It provides theoretical basis for its application in practical research. Materials and methods: in the follow-up study of nutrition intervention trial in the general population of Linxian County, the relationship between fresh fruit intake and long-term death risk of esophageal cancer was evaluated thoroughly by using the commonly used competitive risk analysis method. To explore the impact of competition events on the long-term mortality risk of esophageal cancer during the follow-up period, to estimate the baseline death risk of esophageal cancer, to evaluate the advantages and disadvantages of different competitive risk analysis methods, and to put forward the general strategy of competitive risk analysis. Results: 1, when there is a competitive risk bias, the relative risk function and the relative risk ratio can fully evaluate the effect of the competition event on the risk of the outcome event. The virtual observation method can be used to estimate the baseline risk of the target outcome event, and compared with the results of cause-specific risk regression analysis, The cumulative risk regression analysis was used to estimate the correlation between the factors studied and the target outcome events more accurately. The results of cause-specific risk regression analysis showed that subjects who ate fresh fruit more than once a week were more likely to have a lower long-term risk of death from esophageal cancer than those who did not eat fresh fruit. Moreover, the protective effect did not increase with the increase of fresh fruit intake (PIM trend 0.444), and the cumulative risk regression analysis showed that the intake of fresh fruit might not reduce the long-term mortality risk of esophageal cancer. These results suggest that the existence of competitive risk bias exaggerates the association between fresh fruit intake and long-term death risk of esophageal cancer. The results of competitive risk analysis showed that the effect of competition events on the death risk of esophageal cancer increased with the prolongation of follow-up time, especially the death of cardiovascular and cerebrovascular diseases. The expression of competitive risk effect was different in different fresh fruit intake frequency groups. Further analysis showed that the baseline mortality risk of esophageal cancer was 1.39 in this cohort and 1.62after adjusting for the intake frequency of fresh fruit, indicating whether the intake of fresh fruit can reduce the risk of death in this cohort. Risk of long-term death from esophageal cancer in the population. Conclusion: in the medical follow-up study, cumulative risk regression analysis was recommended to evaluate the association between the factors studied and the disease, and the relative risk function and / or relative risk ratio were used to describe the impact of competitive events on the risk of disease. Improving the decidibility of the study results. 2. The intake of fresh fruit may not reduce the long-term mortality risk of esophageal cancer in the follow-up cohort of nutrition intervention trials in Linxian County.
【学位授予单位】：北京协和医学院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R735.1
本文编号：2124830
[Abstract]:Objective: to evaluate the current commonly used methods of competitive risk analysis, to explore the impact of competition events on the risk of target outcome events, to compare the advantages and disadvantages of different analysis methods, and to put forward the general strategy of competitive risk analysis. It provides theoretical basis for its application in practical research. Materials and methods: in the follow-up study of nutrition intervention trial in the general population of Linxian County, the relationship between fresh fruit intake and long-term death risk of esophageal cancer was evaluated thoroughly by using the commonly used competitive risk analysis method. To explore the impact of competition events on the long-term mortality risk of esophageal cancer during the follow-up period, to estimate the baseline death risk of esophageal cancer, to evaluate the advantages and disadvantages of different competitive risk analysis methods, and to put forward the general strategy of competitive risk analysis. Results: 1, when there is a competitive risk bias, the relative risk function and the relative risk ratio can fully evaluate the effect of the competition event on the risk of the outcome event. The virtual observation method can be used to estimate the baseline risk of the target outcome event, and compared with the results of cause-specific risk regression analysis, The cumulative risk regression analysis was used to estimate the correlation between the factors studied and the target outcome events more accurately. The results of cause-specific risk regression analysis showed that subjects who ate fresh fruit more than once a week were more likely to have a lower long-term risk of death from esophageal cancer than those who did not eat fresh fruit. Moreover, the protective effect did not increase with the increase of fresh fruit intake (PIM trend 0.444), and the cumulative risk regression analysis showed that the intake of fresh fruit might not reduce the long-term mortality risk of esophageal cancer. These results suggest that the existence of competitive risk bias exaggerates the association between fresh fruit intake and long-term death risk of esophageal cancer. The results of competitive risk analysis showed that the effect of competition events on the death risk of esophageal cancer increased with the prolongation of follow-up time, especially the death of cardiovascular and cerebrovascular diseases. The expression of competitive risk effect was different in different fresh fruit intake frequency groups. Further analysis showed that the baseline mortality risk of esophageal cancer was 1.39 in this cohort and 1.62after adjusting for the intake frequency of fresh fruit, indicating whether the intake of fresh fruit can reduce the risk of death in this cohort. Risk of long-term death from esophageal cancer in the population. Conclusion: in the medical follow-up study, cumulative risk regression analysis was recommended to evaluate the association between the factors studied and the disease, and the relative risk function and / or relative risk ratio were used to describe the impact of competitive events on the risk of disease. Improving the decidibility of the study results. 2. The intake of fresh fruit may not reduce the long-term mortality risk of esophageal cancer in the follow-up cohort of nutrition intervention trials in Linxian County.
【学位授予单位】：北京协和医学院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R735.1
【参考文献】
相关期刊论文 前9条
1 杨召;王少明;梁赫;乔友林;范金虎;;竞争风险型数据统计分析理论研究进展[J];中国卫生统计;2016年06期
2 李海彬;李霞;王安心;陶丽新;刘龙;陈斯鹏;郭晋;郭秀花;;竞争风险模型及其在Stata软件实现[J];中国卫生统计;2016年05期
3 杨召;王少明;梁赫;于佩;范金虎;乔友林;;新鲜水果摄入可能降低食管癌长期死亡风险[J];中国肿瘤临床;2016年18期
4 肖媛媛;许传志;赵耐青;;常用生存分析模型及其对时依性协变量效应的估计方法[J];中国卫生统计;2016年03期
5 戚晴;王洪源;任倩;纪立农;;竞争风险模型在2型糖尿病治疗临床试验中的应用[J];中华临床医师杂志(电子版);2012年12期
6 陈征;Nakamura Tsuyoshi;;基于竞争风险理论和概要型数据的病死率估计模型[J];中国卫生统计;2010年03期
7 江一涛;胡海兰;魏巧玲;方亚;;竞争风险模型的发展与应用[J];中国卫生统计;2009年04期
8 陶庄;;使用R软件分析竞争风险模型简明攻略[J];中国卫生统计;2008年06期
9 余红梅;多状态模型简介[J];中国卫生统计;2005年03期
相关硕士学位论文 前1条
1 宋艳龙;竞争风险模型在阿尔茨海默病转归研究中的应用[D];山西医科大学;2014年
，本文编号：2124830
本文链接：https://www.wllwen.com/yixuelunwen/zlx/2124830.html
