糖尿病临床试验中低血糖事件的统计分析

发布时间：2018-01-12 22:36

本文关键词：糖尿病临床试验中低血糖事件的统计分析　出处：《华东师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：糖尿病是一组以持续高血糖为特征的代谢性疾病。糖尿病治疗的主要手段就是要降低血糖,但是过低的血糖也是不好的。因此监测服用药物后低血糖事件的发生,是很重要的一个药物安全分析终点。特别是当受试者因为某种原因多服用了药物,那么会不会导致更多低血糖事件,甚至是严重低血糖事件的发生呢?国外的某个药厂就此为研究目的进行了一个分为两个周期的交换设计(Crossover)临床试验。在这个试验中,大约64名受试者被分为两组,然后研究者监测在每个周期服用双倍剂量药物后84小时内发生低血糖事件的情况。我们在这个研究中涉及到的数据分析指标有低血糖事件的发生与否,发生比率,生存时间和复发事件生存时间。通过对这些指标的统计分析,我们想要探索的目标是试验药是否比对照药在安全性上更好。我们首先从独立性的研究开始,用到的方法有卡方检验,Odds比等。而逻辑回归模型的引入,则让我们同时可以考虑其他一些协变量的影响。以上这些方法针对的是发生与否。而负二项分布回归模型是对发生比率的分析。同时本文中还运用了生存分析,Anderson-Gill模型等一些方法,对生存时间和复发事件生存时间进行了分析。我们的统计分析结果符合我们的预期,试验药比对照药在安全性上更好。同时由于我们的分析基于不用的分析指标,运用了不同的统计分析方法,得到了比较一致的结果,因此我们有比较大的把握认为我们的分析结果是稳定可靠的。
[Abstract]:Diabetes is a metabolic disease characterized by sustained high blood glucose. The main means for the treatment of diabetes is to reduce blood glucose, but low blood sugar is not good. So monitoring the incidence of hypoglycemic events after taking the drug, a drug safety end point analysis is very important. Especially when the subjects for some reason take the medicine, it will not lead to more hypoglycemic events, even severe hypoglycemia events? A foreign pharmaceutical companies this is a purpose for the study was divided into two periods of exchange (Crossover) clinical trials. In this experiment, approximately 64 subjects were divided into two the researchers then monitoring group, taking hypoglycemic events occurred 84 hours after the double dose of the drug in each cycle. We involved in the study of data analysis indicators of hypoglycemia in hair The ratio of students or not, the survival time and survival time of recurrent events. Through the statistic analysis of these norms, we want to explore the goal was to test whether the drug is better than that of the control drug in security. We started from the independent research, use methods of chi square test, Odds ratio is introduced. And the logistic regression model, so we can also consider the impact of some of the other covariates. Based on the above method is occurring or not. And two negative binomial regression model is the analysis on the occurrence rate. At the same time this paper has been used in survival analysis, Anderson-Gill model and some other methods, the survival time and recurrent events the survival time was analyzed. The statistical analysis of our results are consistent with our expectations, the test drug better than control drug safety. At the same time because our analysis index not based on the use of different The statistical analysis method has obtained a relatively consistent result, so we have a relatively large grasp that our analysis results are stable and reliable.

【学位授予单位】：华东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O212.1

【相似文献】