生物医学统计视域下的因果关系探索

Exploration of Causality Based on Biomedicine Statistics

生物医学统计技术依据经验因果性与机制因果性,经验因果性作为使用统计技术的信念,机制因果性则表现为一种动态进程。使用统计技术表征因果机制的局限是群体的未知背景要素问题。群体的未知背景要素在数量上过于庞大,在属性上也无法完全获知。表征因果机制的关键点是以因果路径分析克服未知背景要素的困难,以概率统计表征不同因果路径的效应。统计技术运用相关性概念替代因果性的理念。以超重与高血压之间的因果关系为例进行分析,反事实条件模型表征了两者之间的因果路径。统计技术分析的因果关系在基因科学实践中得到了验证。

Statistical analysis of biomedicine depends on experience-based causality and mechanistic causality.Experience-based causality is the belief of statistical analysis,and mechanistic causality is represented through the dynamic progress.The use of statistical analysis characterization causal mechanism is subject to the unknown background factors,which are large in number and partly accessible to their properties.The key point of characterization causal mechanism is to analyze and conquer the difficulty of the unknown background factors with a causality path,and count and characterize effects of different causality paths.The concept of the correlation is substituted for the concept of causality.This paper characterizes the causality paths with counterfactual condition model tables,which takes the“causality between overweight and blood pressure”for an example.The causality in statistical analysis is verified in the practice of genetic science.