Skip to content

Modeling mortality

To grasp the true impact of COVID-19 on our societies, we need to know the effect of the pandemic on mortality. In other words, we need to know how many deaths can be attributed to the virus, directly and indirectly.

It is already popular to visualize mortality in order to gauge the impact of the pandemic in different countries. You might have seen at least some of these graphs and websites: FTEconomistOur World in DataCBSEFTACDCEUROSTAT, and EUROMOMO. But estimating the impact of COVID-19 on mortality is also controversial, with people either misunderstanding or distrusting the way in which the impact is measured and assessed.

That’s why, I put together a step-by-step guide about how we can go about estimating the impact of COVID-19 on mortality. In the guide, I build a large number of statistical models that we can use to predict expected mortality in 2020. The complexity of the models ranges from the simplest, based only on weekly averages from past years, to what is currently the state of the art.

But this is not all. What I also do is review the predictive performance of all of these models, so that we know which ones work best. I run the models on publicly available data from the Netherlands, I use only the open software R, and I share the code, so anyone can check, replicate and extend the exercise.

The guide is available here:

I hope this guide will provide some transparency about how expected mortality is and can be estimated from data and how it can be used to identify the effect of the pandemic. Feel free to try different estimation strategies, adapt the code to data from different countries, or use different measures to compare the performance of the models.

Published inData visualizationObservational studiesR

Be First to Comment

Leave a Reply

Your email address will not be published.