论文部分内容阅读
目的:建立一种统计学模型来评价临床不良事件与治疗间的关系。方法:将交叉试验的治疗期与间歇期分开,记录每一个阶段不良事件发生的次数和持续天数,累积各受试者每一阶段发生不良事件的总时间,用日/人表示。以泊松分布法计数发生事件反应的次数及其持续时间;以普通线性混合模型估计各不同处理间不良事件的发生率;以广义估计方程估计各交叉试验组间的相关性。结果:不良事件的发生率不仅受发生次数的影响,也受其持续时间和受试者耗于不良事件(日/人)总时间的影响。结论:本研究提供了一个实用的评价各种不良事件发生率的方法,此方法比单纯考虑不良事件的发生次数具有更少的偏性。
Objective: To establish a statistical model to evaluate the relationship between clinical adverse events and therapies. Methods: The treatment period and interim period of the cross-test were separated. The number of occurrence of adverse events and the number of days of persistence in each phase were recorded. The total time for accumulating the occurrence of adverse events in each phase of each subject was expressed as day / person. The Poisson distribution was used to count the number of events and the duration of events. The linear mixed model was used to estimate the incidence of adverse events among different treatments. The generalized estimation equation was used to estimate the correlation between the cross-over groups. RESULTS: The incidence of adverse events was affected not only by the number of occurrences but also by their duration and subject’s total adverse event (day / person) duration. CONCLUSIONS: This study provides a useful method of assessing the incidence of various adverse events and has less bias than simply considering the number of adverse events.