Miettinen O S, Wang J D
Am J Epidemiol. 1981 Jul;114(1):144-8. doi: 10.1093/oxfordjournals.aje.a113161.
The conditions that are generally considered to justify the use of the proportionate mortality ratio (PMR) to compare exposed and nonexposed populations also permit the computation of the ratio of mortality odds between the cause (of death) of interest and the "other" causes considered. This mortality odds ratio (MOR) comparing the exposed with the nonexposed equals the exposure odds ratio comparing deaths from the cause of interest with those from the auxiliary ("other") causes, and, in contrast to the PMR, it can be interpreted as the observed-to-expected ratio or the standardized mortality ratio on the assumption that the mortality rate for the auxiliary causes is unrelated to the exposure. Related to this, the MOR is free of the arbitrary element in the PMR, which is dependent on the size of the auxiliary-causes domain. These properties make the proposed measure superior to the PMR statistic. Consideration of the proposed statistic for "PMR data" also underscores the need to view "PMR studies" as case-referent studies, and this has important implications for the choice of the auxiliary causes (reference series).
通常认为使用比例死亡率（PMR）来比较暴露人群和非暴露人群的情况，也允许计算感兴趣的死因与所考虑的“其他”死因之间的死亡比值比。这种比较暴露组与非暴露组的死亡比值比（MOR）等于比较感兴趣死因的死亡与辅助（“其他”）死因的死亡的暴露比值比，并且与PMR不同的是，在假设辅助死因的死亡率与暴露无关的情况下，它可以被解释为观察到的与预期的比值或标准化死亡率。与此相关的是，MOR没有PMR中的任意元素，PMR依赖于辅助死因范围的大小。这些特性使得所提出的测量方法优于PMR统计量。对“PMR数据”的建议统计量的考虑也强调了将“PMR研究”视为病例对照研究的必要性，这对辅助死因（参考系列）的选择具有重要意义。
