Calculation of week-specific age-standardized death rates from STMF data on mortality by broad age intervals

The Short-Term Mortality Fluctuations (STMF) data series provides an opportunity for analysis of intra-annual excess mortality, in particular, human losses due to the COVID-19 pandemic. Nevertheless, the STMF has a limitation caused by the nature of the collected original weekly death counts. In many countries, weekly death counts are available only by broad age groups or/and are too small and shaky. Moreover, the original age scales somewhat vary by country. Thus, the STMF data file presents weekly deaths and death rates by broad age intervals. This simplifies the usage of the STMF and helps to conduct analyses but limits the comparability of results across countries and time. The comparisons may be biased due to differences between the population age composition. This study addresses the problem by providing a method for the estimation of week-specific standardized death rates (SDRs) that combines the aggregated weekly mortality data with detailed annual data on mortality and population. This allows deriving annual transition coefficients for the transformation of crude death rates into SDRs. We show that the derived SDRs approximate well exact SDRs across time and countries.