Zeighami E A, Morris M D
Am J Epidemiol. 1983 Jan;117(1):90-7. doi: 10.1093/oxfordjournals.aje.a113520.
The standardized proportionate mortality ratio (SPMR) is shown to be a summary measure which is a weighted average of age-specific proportionate mortality ratios which uses an internally derived set of standard weights. As with any summary measure, the SPMR is only meaningful when the stratum-specific values can be judged to represent a common value. When stratum-specific values are not homogeneous, the use of the stratum-specific values themselves is more appropriate than the use of a summary measure. A test for homogeneity of the age-specific proportionate mortality ratios is presented. The test also provides an estimate of the common value and its variance, when that single parameter can be presumed to exist. A procedure of external standardization is presented which uses weights derived from the standard population.
标准化比例死亡率(SPMR)被证明是一种汇总指标,它是年龄别比例死亡率的加权平均值,使用的是内部推导得出的一组标准权重。与任何汇总指标一样,只有当特定分层的值能够被判定代表一个共同值时,SPMR才有意义。当特定分层的值不均匀时,使用特定分层的值本身比使用汇总指标更合适。本文提出了一种针对年龄别比例死亡率同质性的检验方法。当可以假定存在单一参数时,该检验还能提供对共同值及其方差的估计。本文还提出了一种外部标准化程序,该程序使用从标准人群得出的权重。