In the studies below we analyse the dangers and pitfalls of using the standard measures for excess deaths and how they can lead to misguided policy decisions. In addition, we propose an alternative way of computing excess deaths based upon computing excess death rates instead of excess deaths.
We show that using prior N-year average of deaths as baseline (method 1) is inappropriate in estimating excess mortality. The volatility of the changes in deaths as well as the problem of age group populations oscillating over time, makes this measurement inaccurate, leading to severe biases. This could then lead to misguided assessments and consequently erroneous policy actions.
Death rates tend to decline over time, for a given population age cohort, as living conditions have generally been improving since after the Second World War (with some exceptions). Estimating excess mortality by computing excess death rates obtains a much clearer signal, this is, a much more accurate estimate for excess mortality. This is because changes in death rates do not trend upwards and downwards over time and are much less volatile from year-to-year than changes in deaths.
Consequently, by using the death rate in 2019 (method 2A) as a baseline to estimate excess death rates in 2020 to 2022 will tend to be a lower limit on excess deaths. A more realistic assessment is to compute excess death rates in 2020 to 2022 relative to a baseline that adjusts 2019 death rates for average yearly rates of improvement in death rates (method 2B), or, use a baseline that is the continuation of a previous trend in excess deaths (method 2C).
PHINANCE TECHNOLOGIES