Why counting coronavirus deaths is not an exact science | Gianluca Baio and Marta Blangiardo

It's the data that gets all the headlines, but the two main sources of information in the UK have their own limits

"Nothing can be said to be certain, except death and taxes." Right? Well, not quite. As many people are discovering during the Covid-19 crisis, even something as seemingly straightforward as the count of people who have died might not be as robust as to be taken at face value. In the UK, distinctions between the data from the Department of Health and Social Care (DHSC) and the Office for National Statistics (ONS) have led to talk of whether deaths are being underestimated and even whether the figures can be trusted.

The pandemic highlights several difficulties in counting, reporting and modelling mortality data. Arguably, the most important problem is the "denominator" - what is the actual number of people who are infected by the virus? This is virtually impossible to determine, except perhaps in the unlikely scenario of real-time, continuous, population-wide testing. The absence of this figure creates problems when rescaling the outcomes - for example, deaths - to the number of people at risk, which is not known with precision. And because we cannot know for certain the total number of infected, international comparisons also become tricky: does Germany have more cases than the UK because it tests more and reports the results more systematically?