Occupational mortality indices

Summary This paper demonstrates the danger of excessive reliance on an occupational mortality index such as the Standardised Mortality Ratio. An Index can be considered as a valid estimate of a group’s relative mortality only if the (age) specific mortality ratios which constitute the index are not significantly different. The homogeneity of these (age) specific mortality ratios should therefore be tested whenever possible The Standardised Mortality Ratio is shown to be equivalent to the maximum likelihood estimator of a common (age) specific mortality ratio. It follows that the standard error of the Standardised Mortality Ratio is smaller than the standard error of any other mortality ratio. The relevance of these considerations to comparisons of social class mortality is discussed and illustrated by the analysis of mortality from four separate causes of death and from all causes of death combined.