Summary

Measuring the impact of a public health crisis in terms of mortality might seem a straightforward method to quantify its effect on the population because deaths are much more easily registered compared to other health outcomes. However, despite the intuitive appeal of this path, it is far from obvious how to best operationalize it, and all the most used methods have drawbacks that should be kept in mind. Especially during the COVID-19 pandemic, the major routes that have been considered are cause-specific death counts (and related measures such as case fatality rates), excess deaths estimates, and life expectancy decline. All the considered approaches have limitations: Cause-specific deaths are often subject to undercount or overcount issues with significant differences both between and within countries, excess deaths estimates may strongly depend on the baseline (there are several methods to estimate it), and life expectancy drop estimates (or estimates of years of life lost) also depend on the reference level used, which can vary substantially across countries. More generally, the issues of available data quality and standardization of age structure should be taken into proper account. Thus, the choice of which approach is worth using depends on the characteristics of the crisis that need to be evaluated and the type and quality of data available. Interestingly, the three approaches can also be combined so that some of their limitations can be mitigated.