论文部分内容阅读
为解决在职业暴露跟踪调查及重复测量研究中,健康工人效应会使研究结果出现偏差的问题,采用新型的边缘结构模型(MSM)对相关数据进行处理。通过构造与事实相反的虚构样本,同时处理不同情况下的数据删失问题,以寻找跟踪调查的中间结果作为时变混杂变量,进而计算稳定反概率处理权重(IPTW)以修正健康工人效应造成的统计结果偏差,并用MSM来处理实际样本数据。结果表明:MSM可以通过使用稳定IPTW在一定程度上减少职业暴露研究数据分析过程中健康工人效应对研究的影响,比以往的数据分析方法更能准确地反映重复测量跟踪调查研究中职业暴露水平与职业危害间的关系。
In order to solve the problem of deviations caused by the effect of health workers in the follow-up survey of occupational exposure and repeated measures, a new type of edge structure model (MSM) was used to process the relevant data. By constructing a fictitious sample that is contrary to the facts, we deal with the problem of data censorship in different situations at the same time to find the intermediate results of the follow-up survey as the time-varying mixed variables and then calculate the stable inverse probability processing weight (IPTW) to correct the effect of health workers Statistical results deviate and use MSM to process the actual sample data. The results show that MSM can reduce the impact of health workers’ effects on occupational exposure research data analysis to a certain extent by using stable IPTW, reflect the occupational exposure level in repeated measures follow-up survey more accurately than previous data analysis methods and Occupational hazards between the relationship.