Correlation: Confinements and mortality
In response to this pandemic, many countries have opted for extreme measures. General lockdown was introduced on the grounds that it was a way of decreasing transmission and hence mortality from disease (directly or through disruption of health systems).
Its side effects are substantial and can affect the increase of mortality (even in the short to medium term).
An analysis of the correlation between this type of measure and mortality is required. We have opted for general mortality for several reasons:
- I have included all possible impacts – positive and negative;
- There are differences in the tracing and classification of “death by Covid” between countries;
- There are very particular criteria in the attribution of the classification: death “by Covid”;
We have also included a reference to the seasons because of the impact they generally have on mortality associated with pandemics (although Z-score has already incorporated it).
Many analyses have been made using time-limited correlations or isolated cases. Such partial analyses often lead to “cherry picking”, i.e. the choice of partial information to confirm a given perspective.
Rather than drawing conclusions for each country, we present the observation of all countries -included in the EuroMOMO– in terms of mortality, using the variables (mandatory confinement, recommended confinement, no recommendations and seasonality).
In order to have a standardization and independent classification of the measurements, we used the Coronavirus Government Response Tracker from Oxford University (Our Word in Data).
Cautionary note: We are observing the correlation between variables, confinement and (general) mortality. However, correlation is not necessarily causality. There are other factors that may interfere with general mortality.
The chart for Portugal has been rectified due to an inaccuracy detected in the data taken from Our World In Data.
*Text written in collaboration with Eduardo Pinto Leite