核保医学研究中常见统计偏倚与混杂的控制

摘要

统计偏倚和混杂是核保医学研究中非常重要的影响条件，如果将各类风险因素不加选择地罗列其中，就会不可避免地引入偏倚，特别是混杂偏倚的引入会影响研究因素与结局之间真正的联系，但这些问题并不能利用统计方法解决，需要在研究设计阶段进行充分考量。%The statistical bias and confounding are the most important influence conditions in the underwriting medical re-search, the bias can be inevitably introduced if various risk factors are indiscriminately listed in it, especially the introduc-tion of confounding and bias can affect the real correlation between research factors and outcomes, but the statistical method fails to solve these problems, and they need to be fully considered in the research and design stage.