病因统计推断中的因果关系研究

摘要

基于观察研究的病因推断是一个医学难题,由于疾病产生的原因往往混杂着社会、心理、环境和生理等复杂因素,因而因果关系并非简单就能证明出来.统计模型在病因推断中有其成立的先验假设条件,有些条件是不可经验证伪的.讨论统计推断中因果关系的含义,揭示因果统计推断背后隐藏的假设条件,从而看到病因统计推断的局限性.%Etiological inference based on observation study is a hard problem in medical science. It is uneasy to prove causality, because the causes of disease are often involved with the social, psychological, environmental and physiological factors. Statistical models carry out with prior assumptions, some of which are not empirical falsifi-able. This paper aims to discuss the concept of causality in statistical inference and reveal some assumptions rooted in causal statistical inference in order to see the limitations of statistical inference for etiology.