Confusingly, COVID-19 death rates seem to differ widely between countries and even by different newspapers in the UK. This post aims to explain how these statistics are calculated and potential explanations for the different rates we’re seeing - giving you the tools to be able to interpret for yourself how fatal COVID-19 is.
Let’s start from the beginning: what does mortality and fatality rate actually mean?
We used mortality and fatality death rate instead of just ‘the number of people who have died’ because the number of deaths depends on the period of time you’re measuring in and the overall size of the population. Mortality rate describes the proportion of deaths as related to the whole population, whilst fatality rate is the proportion of deaths where everyone is infected. A couple of examples are enough to show how different they are:
Mortality rate depends on how many people get infected and how deadly the disease is, whilst fatality rate is just how deadly the disease is once you know you have it.
In these two examples, we know exactly all our important numbers: how many people there are in our population, how many people have the disease and how many people have died from the disease. In real life, these numbers are much harder to pin down - particularly a) how many people have the disease and b) how many people die from the disease.
a) How many people are infected with COVID-19?
We can only know the answer to this question if we test everyone for COVID-19. Currently in the UK you are tested if you go to hospital with COVID-19-like symptoms or if you’re a key worker. Due to asymptotic infection, it’s very possible there are lots of people with no symptoms walking around that aren’t included in our stats. This means that the differences between countries in how many people they are testing are likely to account for a lot of the different fatality rates. For example more widespread testing in Germany could result in apparently lower death rates - if you’re dividing the same number of deaths by a bigger number of people infected, you’ll end up with a smaller fraction.
There are many reasons for why it is difficult to test everyone for COVID-19, including logistics, lab capacity and how often you test everyone. This is one of the reasons why there is a demand for antibody tests, which could be used to tell if someone had been previously infected.
b) How many people have died from COVID-19?
The people most vulnerable to COVID-19 usually have other underlying health conditions such as respiratory or cardiovascular disease, or cancers. So if they die in hospital, how can the doctors tell whether it was due to COVID-19 or their other conditions? This question becomes even harder if the patient hasn’t been tested for COVID-19 - do you predict they had COVID-19 because of their symptoms or not? You can see these differences even from different organisations in the UK - GOV.UK for example only counts COVID-19 deaths that occur in hospital where patients had a positive test result for the disease (because this was agreed as a European standard way to report deaths so that these statistics have some consistency for comparison between European countries), whilst the ONS (Office for National Statistics) include all deaths where COVID-19 was suspected, and include non-hospital deaths. Most countries are reporting deaths in hospitals of people who have tested positive for COVID-19 as caused by COVID-19, even if it’s possible that they have died from other causes.
Just to make it even more complicated, these rates are constantly changing - each day more people are getting tested or dying from the disease, changing all the values. Also, it’s worth bearing in mind that rates can be presented as a percentage (%) or per 1 million or thousand people. Flu deaths are measured in ‘excess deaths’, the number of extra deaths we are seeing compared to what we normally see at this time of the year. We can’t know for certain that all the excess deaths we are currently seeing are due to COVID-19, and we can’t accurately measure excess deaths whilst the pandemic is occurring, but it is one of the measures helping scientists model and understand the severity of the disease.
Unfortunately, “how fatal is COVID-19?” is not a simple question to answer. There are various values, from fatality rate to excess deaths, that can be calculated - but these values don’t mean exactly the same thing. What’s more is that these values are only as reliable as the data used to calculate them, namely the numbers of infected people and number of people dying from COVID-19. Hopefully this background on where these numbers come from will help you navigate the statistics which are constantly being collected and improved in an effort to better inform our response to coronavirus.