Estimating local mortality tables for small areas: An application using Belgian sub-arrondissements
Standardised mortality ratios (SMR) may give a good estimate of the relative level of mortality in a local area, and its relation to local social conditions, but if we wish to understand changes in the age distribution of mortality as mortality declines, we need an estimate of the local mortality curve. Such fine detail can be elusive when examining small populations for which the number of people in each age group is small, the number of deaths minuscule, and estimation errors are large. A possible solution to this problem is to estimate age-specific mortality rates simultaneously for all the subunits of a particular country, using the reported number of deaths, by age and sex, for each unit as the input data. The national mortality rates then serve as a model from which local deviations, by age and sex, are estimated, on the basis of overall mortality (SMR) and local social conditions. We demonstrate this approach using data from 87 sub-national units in Belgium to construct local-level life tables, using a multilevel model with the local sex- and age-specific cells as units, nested within sex-age groups and regional units at the second level. The results indicate that life expectancy is closely related to SMR, but the specific shape of the mortality curve, in particular the range over which mid-life mortality is low and the age at which mortality begins to rise into senescence, varies by level of mortality and social conditions.